IMSS21: 11TH INTERNATIONAL SYMPOSIUM ON INTELLIGENT MANUFACTURING AND SERVICE SYSTEMS
PROGRAM FOR FRIDAY, MAY 28TH
Days:
previous day
all days

View: session overviewtalk overview

09:00-10:40 Session 5A: Parallel Session
09:00
A Case Study of Shape Optimization Using Grasshopper Optimization Algorithm

ABSTRACT. Structural optimization is a popular topic in today's engineering and industry to reduce costs and obtain more ideal designs. In the structural optimization problem, the lightest design under conditions is investigated. Increasing studies in recent years have proven the success of meta-heuristic optimization algorithms in structural optimization problems. Unlike the traditional method, Meta-heuristic methods use stochastic methods and do not need derivative information of the problem. This makes them more flexible and easy to use. In this study, a solid part's shape optimization is performed by using the Grasshopper Optimization Algorithm (GOA). The design values and results found by the algorithm are discussed.

09:20
An Application of TSP Maximization with Grey Wolf Optimizer: Dartboard Design Problem

ABSTRACT. Dart game is one of the most played games in the British Islands, with approximately 6,000,000 entrants. Its roots date back to ancient times, and it is said that the travelers on the Mayflower played it in 1620. Brian Gamlin invented the new scoring system in 1896. Researchers discussed various dartboard designs after researching the problem of optimally positioning the numbers around over a dartboard as either a Traveling Salesman Problem (TSP) and a Quadratic Assignment Problem (QAP). However, Selkirk (1976) depicts the most traditional form of the dartboard in the literature. In this study, optimal dartboard design problem implemented with Grey Wolf Optimizer Algorithm and MATLAB software evaluates based on squares of adjacent number differences. S. Mirjalili et al. (2014) developed the Grey Wolf Optimizer (GWO), a swarm-based optimization and meta-heuristic algorithm that has been used by several researchers to solve several optimization problems. The Grey Wolf Optimizer algorithm was selected to solve the optimal dartboard design problem because it was observed to be appropriate in the application of combinatorial optimization problems according to a literature review. Within the scope of this study, it is aimed to create an optimal classic dartboard design based on the maximum global best solution values.

09:40
META-HEURISTICAL CONSTRAINED OPTIMIZATION BASED ON A MECHANICAL DESIGN PROBLEM

ABSTRACT. Mechanical design optimization is a crucial part of engineering and manufacturing problems, according to the literature. Dealing with constraints is amongst the most challenging parts of this form of problem. Many studies have been presented to concern with infeasible solutions in process optimization, especially those based on genetic algorithms. The application step is carried out in this study with a chosen problem that is investigated by a Group Search Optimizer (iGSO) for resolving mechanical design optimization problems and an algorithm that uses both subsets and cooperative evolutionary strategies to enhance swarm intelligence capabilities and optimization accuracy. The selected pressure vessel design from this study aims to minimize the total cost of materials, forming and welding of the pressure vessel. By this aim, the pressure vessel design optimization problem was implemented with Simulated Annealing and Great Deluge Algorithms via MATLAB R2020b software. The parameters and constraints were specified in two distinct functions to compute the fitness function using an inverse tangent approach and even to obtain a penalty function that decreases the fitness values of infeasible solutions in accordance with the degree of constraint violation. Since it has a tiny area and performs in a restricted period, the Simulated Annealing Algorithm offered a reasonable result despite violating any constraints, according to MATLAB outputs.

10:00
A two-echelon pharmaceutical supply chain optimization via genetic algorithm
PRESENTER: Elif Yıldırım

ABSTRACT. Two-echelon supply chain model consists of two separate components that have diverse objectives. In this study, a supply chain is modelled as a two-echelon supply chain consisting of one supplier, one retailer and one product at a pharmaceutical supply chain. The purpose of optimization is to maximize the total profit generated during sales and distribution of the product. Sales take place at the pharmaceutical retailer and the demand encountered is stochastic, and order periods are determined by the frequency of visits by the drug supplier. Considering this visit frequency, the pharmaceutical retailer uses a periodic review inventory system. For the retailer, the decision variable is the safety factor which is determined according to the selected order level and determines the pharmaceutical supplier profitability by affecting the sales volume of the Pharmaceutical Supply Chain. The problems optimized with two different scenarios and two different models which are the traditional supply chain model, and the two-echelon supply chain model. Heuristic model of these problems provides more profitability for both echelons. With this study it is understood that genetic algorithm can be used for two echelon pharmaceutical supply chain and gives better results than mathematical model of the same problem.

10:20
A Mathematical Model for Part/Material Supply Problem in Just-in-Time Environment
PRESENTER: Gül Gündüz

ABSTRACT. We study a just-in-time production system in the automotive industry, in which car parts are delivered to the production line from a central parts store, using a set of dollies. This paper introduces a mathematical model for the part supply plan that aims to calculate optimal routes and schedules for the dollies. The proposed mathematical model is developed as a mixed-integer-programming model, and the validity of the proposed approach are examined in a comprehensive computational study using real-life information.

09:00-10:40 Session 5B: Parallel Session
09:00
Evaluation of the challenges in Industry 4.0 Applications in Sustainable Supply Chain Management in Developing Economies
PRESENTER: Feyza GÖr

ABSTRACT. Industry 4.0 is a new concept that is necessary to ensure sustainable supply chain management in the manufacturing industry and also affects the entire production system by transforming production processes such as design, production and delivery. In this study, it is aimed to define the main difficulties in Industry 4.0 applications and determine the priorities of these difficulties for sustainability supply chain of in developing companies in Kayseri Organized Industrial Zone. A framework consisting of a two-level model is presented to create a quantitative measure of the challenges in Industry 4.0 applications. At the basic level, a fuzzy artificial decision was created based on data from the companies participating in the survey. At the upper level, a fuzzy multi-criteria decision-making model was used to measure the difficulties in Industry 4.0 applications. This study will be useful to evaluate the challenges of Industry 4.0 implementations in manufacturing companies for a sustainable supply chain and will also provide adequate understanding of industry 4.0 for managers and organizations.

09:20
A Conceptual Model of Solid Waste Management System for Sakarya Province
PRESENTER: Safiye Turgay

ABSTRACT. With the developing economy, increasing population, and urbanization rate, urban solid waste management gains great importance. In this study, a conceptual model of municipal solid waste management is presented. Parameters, variables, and constraints affecting the multi-purpose decision-making model structure were determined and modeled. Since the developed model is a general model structure, it can easily apply to other cities. In the next study, multi-purpose optimization will be discussed as a problem solution considering different scenario situations. With rapid urbanization and increasing waste rates with developing technology, environmental damage and human health are negatively affected. The main purpose of modeling is to minimize the total cost and negative environmental effects. In this context, its maximum value is also important in the amount of fertilizer and energy obtained from material recovery and waste. Therefore, it is aimed that the facility to be established will operate at minimum cost and provide the highest efficiency.

09:40
A Novel Linguistic Approach for Technology Assessment in Smart Agriculture
PRESENTER: Deniz Uztürk

ABSTRACT. Latterly, science tries to project the state of novel trends related to Industry 4.0 technologies. They are also accepted as the key enablers to reach a circular economy. Scientists search to model and understand the trajectory of developing technologies in the circular economy. Decision-making methodologies are one of the beneficial instruments to understand future topics. Multi-criteria decision-making (MCDM) methods can be helpful to assess and evaluate the developing technologies. Prospective multi-attribute decision-making (PMADM) is a subpart of MCDM approaches. It enables decisions and evaluation based on future aspects. In this paper, the focused area is the smart agriculture subject. The idea of smart agriculture has accelerated with the penetration of Industry 4.0 technologies into traditional agriculture. It is still an emerging and wide-open area thanks to the new application approaches of 4.0 technologies. Consequently, in this paper, a linguistic framework is proposed to evaluate sensor technologies in smart agriculture. Sensor technologies are one of the main components to achieve effective end-to-end control over the whole agrarian system. A group decision-making (GDM) approach is recommended with the novel Measurement of Alternatives and Ranking according to the Compromise solution (MARCOS) method. Plus, the MARCOS method is extended with the 2-tuple linguistic model for the first time to emphasize its ability to deal with multi-granular linguistic information. The 2-Tuple approach increases the accuracy of the linguistic computations without the loss of information and it provides the interpretability of the results. The evaluation criteria for sensor technologies are generated from a comprehensive literature review, and a case study of the proposed methodology is given. The results are analysed, and the influence of dynamic indices is investigated with sensitivity analysis. Plus, a comparative analysis is presented to show the robustness of the 2-tuple integrated MARCOS methodology.

10:00
Design of an Industrial Symbiosis Network System for Agriculture and Animal Husbandry Sectors in Ankara, Turkey

ABSTRACT. Industrial symbiosis is a process which uses wastes or by-products of an industry as raw materials (inputs) for another industry. It creates an interconnected network between different or similar type of industries by exchanging production factors. With the risk of climate change impacts, environmental concerns and scarcity of resources, the need to develop more sustainable systems is urgently necessary. The circular economy is often mentioned as one of the best solutions to support more sustainable systems. Due to increasing population and extended industrialization, landfill need for disposal of wastes and impacts of environmental footprints of the industries are some important environmental problems. Industrial symbiosis reduces these problems by reusing (reproducing, recycling, etc.) waste materials and creates an economic value from waste materials. It extends the knowledge of practical know‐how of how waste management can be transformed into a sustainable and growth oriented business. Currently, Europe has some EU support networks for industrial symbiosis and European Innovation Partnerships. Some examples of these are National Programmes (e.g. NISP (UK)), regional initiatives (e.g. Cleantech Östergötland (Sweden)) and Local initiatives (e.g. Kalundborg in Denmark). Created synergy between companies with symbiosis resulted in eco-industrial parks. An Eco-Industrial Park (EIP) is a community of businesses located together, exchanging material and energy with each other, to achieve sustainability advantages for its participants. Industrial symbiosis networks and eco-industrial parks are also settled extensively in Turkey. The first industrial symbiosis project ‘İskenderun Bay project’ is completed by Technology Development Foundation of Turkey with the partner of International Synergies and financed by BTC Crude Oil Pipeline Company. Several projects are ongoing; for example, Eco-Industrial Park transformation in İzmir, Trakya Industrial Symbiosis program, Bursa, Eskişehir, Bilecik Industrial Symbiosis Program, Gaziantep Industrial Symbiosis project, Eco-efficiency project in Antalya Organized Industrial Zone. On the other hand, Systems Engineering (SE) is an interdisciplinary approach and used for designing successful systems. It focuses on whole of a system with its surrounding environment. We can use a different paradigm to understand and develop a nationwide industrial symbiosis with system engineering approaches. In this study, we consider a sustainable industrial symbiosis network design choosing Ankara as a case study. Participant of the network are selected from agriculture and animal husbandry based firms. In the proposed network, all participants are feeding each other with their waste materials or by-products. By aiming economic and environmental considerations, developed mathematical problem optimizes the network with objectives simultaneously. In the model, all potential linkages between participants are considered. The mathematical optimization model formulates the general network of the proposed case study. It can be transform into single objective problem considering two objectives seperately. There are many alternatives for objective function formulations with respect to decision makers who are the experts are responsible for the strategic plans for the network developments. Generated objective function results can be compared with each other and then; most contributed one can be selected.

10:20
ALTERNATIVE CLEANING COMPANY SELECTION METHOD WITH FUZZY TOPSIS

ABSTRACT. Companies have to make decisions in the face of situations in order to exist in the market. These decisions should be those involving cost saving to the companies while at the same time provide win-win relationships with the suppliers. Supplier selection is one of the most important decision-making processes for companies in terms of cost and benefit. In this process, companies faced with multiple alternatives and criteria find it difficult to choose and achieve the best result. Since the selection of suppliers for the public sector is carried out on a tender basis, although non-cost factors are tried to be taken into account, the weight is mostly left on the supplier company that offers the lowest bid. This may cause poor quality workmanship due to price, even it will not be able to meet factors like the environmental criteria which are also considered in green supplier selection today that may be important in the selection of suppliers. In this study, a proposed alternative selection method of cleaning company was made with the Fuzzy TOPSIS method for a company operating in the public sector. The most important factor in using the Fuzzy method rather than the classical method is the advantage it provides to the solution, since verbal expressions are used during comparison and evaluation. Problem solving was made through 4 supplier companies working in the cleaning business sector and the real names of the suppliers were not used for confidentiality. In solving the problem, 20 sub-criteria based on Economic, Social and Environmental main criteria were used. Satty (1-9) scale for these criteria were transformed into verbal variables developed by Chang (1996) and triangular fuzzy numbers described for these verbal variables, and the selection for the best supplier was tried to be taken with the Fuzzy TOPSIS method.

10:40-10:50Break
10:50-12:30 Session 6A: Parallel Session
10:50
Optimization of Parameters Affecting Hardness and Diffusion Thickness on Boronizing Steels Using Grey Taguchi Method

ABSTRACT. Boronizing treatment is widely used as a coating process to improve the surface properties because of the boride layers’ high melting point, hardness, strength and excellent friction resistance as well as its outstanding chemical stability in some salty solutions, some acidic solutions and oxidative atmospheres. In this study boronizing is applied to different steels such as AISI 5140, AISI 4340 and AISI D2 at temperatures of 850 ˚C, 900 ˚C and 950 ˚C for 2, 4 and 6 hours duration and then is determined their values affecting on the best diffusion thickness and surface hardness values occurring at the same time. Taguchi experiment design method was applied to reduce the number of experiments and thus the cost. However, the traditional Taguchi method allows the improvement of one product characteristic (response), whereas in this study, the characteristics are two. For this, it is appropriate to use the Grey Taguchi method, which is the integrated form of the Taguchi method with the grey relationship analysis. Grey relational analysis achieves to convert the multi response problem into the single response problem. Grey Taguchi method and then ANOVA were applied, and it was found that the effective factors are temperature and time, their values that will make the diffusion thickness and hardness the highest at the same time should be 850 ˚C and 6 hours.

11:10
A bibliometric analysis of concept drift in streaming data from 1980 to 2020

ABSTRACT. Concept drift has become a hot research topic with the increasing popularity of dynamic data. An extensive bibliometric analysis of concept drift is represented in this paper; which embraces the last four decades, namely from 1980 to 2020. Considering the scope and the discipline; the core collection of the Web of Science database is regarded as the resource of this study, and 1,468 publications related to concept drift are obtained. Through the classification of statistics and feature analysis of valid literature data, the bibliometric indicators are revealed such as; prolific countries, regions, institutions, authors. The analyses shed light on the research directions and highly authoritative publications. This paper aims to illustrate the bibliometric analysis method by using VOS viewer software and construct diverse informative graphs to designate the possible directions for more valuable further researches.

11:30
Applications of Artificial Intelligence in Last Mile Delivery Operations: A Literature Review
PRESENTER: Kutay Akdoğan

ABSTRACT. Artificial intelligence (AI) is trending on almost every business field and new researches are being done day by day. Those who cannot adapt to Industry 4.0 will ultimately perish and AI will be an inseparable feature of this challenge. Customer demands increase each day with expanding e-commerce activities and slightest delay meets with customer dissatisfaction. Logistic companies face pressure to deliver increasing number of parcels while performing fast, cost efficient and environmentally friendly. Last mile delivery can be described as the last step for logistic journey which goods are taken from parcel hubs and transported to final consumers. It is known that this step covers majority of the logistics cost hence new solutions must be presented. In an era where the competition is among the supply chains, implementation of AI technologies into the last mile delivery process can improve success of overall operations therefore enhances customer’s quality perception. This paper aims to present an insight on last mile delivery and artificial intelligence usage as it is still considered as one of the biggest hurdle in logistic services. AI utilizations on last mile delivery are increasing considerably and companies are racing to gain technological competitive advantage through AI implementations. Therefore a detailed research is needed to grasp the current situation in the academic literature. To conduct a decent literature review, Scopus database is chosen and "artificial intelligence", "autonomous", "last mile delivery", "last mile distribution" and “last mile transportation" terms are searched across the database without time limitations. It is found that there are 65 documents related to the searched terms and oldest paper was published on 2011. Recorded papers are reviewed, classified and examined to find research gaps; to offer guidance on future researches.

11:50
Effect of Build Orientation on Cross-section Areas of Sliced Layers and Geometrical Accuracy in Selective Laser Melting

ABSTRACT. Additive manufacturing is a manufacturing process that allows the production of complex parts and has many advantages over conventional production methods. However, the pre-processing stage is still time-consuming and open to failure. Build orientation is one of the pre-processing stages, which have a crucial effect on support requirement, build cost, and accuracy of the produced part. In recent years, number of research has been made to optimize build orientation for surface roughness, the requirement of support structures, build time, and cost. For metallic additive manufacturing, limited number of research has been carried out. Selective laser melting is one of the powder bed fusion technology that allows the production of high-performance metallic parts. In the selective laser melting process, some defects may occur due to residual stresses resulting from solidification during the process. Build orientation is important in selective laser melting to ensure proper heat flow throughout to entire structure during the process. After the build orientation is selected, the part slices into layers. Each layer builds on the previous layer, and production carries out. The cross-section areas of these sliced layers depend on the build orientation. This study investigates the effect of cross-section areas on the geometric accuracy of the part. The numerical evaluation shows that the distribution of layers has a significant impact on geometrical accuracy.

12:10
PROCESS IMPROVEMENT WITH LEAN MANUFACTURING TECHNIQUES IN THE CABLE COMPANY
PRESENTER: Seda Pirvan

ABSTRACT. By distributing the amount determined by the customer demands to the production processes, the cables are taken to the intermediate stock area after the cutting process. The intermediate stock area is a transition area between the cutting process and pre-assembly, banding, set-laying, final assembly, ratchet control and finally electrical test processes. In this area, the cables must have 2.5 days stocks depending on the demand. Cables are sent to production processes on a product basis. For this reason, whichever product is in the order to be produced, there are cables to be used in the intermediate stock. The intermediate stock area consists of 28 branches and the length of each arm is 40 cm. When the quantities produced in line with the demand come to the intermediate stock after cutting, they take up space in the stock arm as long as their cross-section. In some products, 28 arms are sufficient and there are idle arms, while in some products, 28 arms may be insufficient and a new intermediate stock area may be needed.

The values​​reflecting the customer demand and the cable types to be used for the desired products are determined. Mathematical models are established by using the operations research technique according to the availability of these cables in the product to be produced and their usage percentages. The installed models are solved with the LINDO program to find out how much of which cable will be produced. Then, mathematical calculations are made over 1 arm by considering the cross-sections of the cables and it is concluded how many suspension arms we need for that product.

All these stages are carried out considering that the company has adopted the lean production technique and works with the minimum stock principle. Lean manufacturing technique principle, quality must be ensured at every stage. When there is a quality-based problem in the cutting process, production may be delayed or when insufficient hanger arms are reached, the cost of new intermediate stock space may be required.

Calculations are made without exceeding the firm's standards, taking into account all the specified stages. As a result of these calculations, if there is insufficiency in the number of suspension arms in order to ensure the continuity of production in other processes, improvement studies will be carried out.

10:50-12:30 Session 6B: Parallel Session
10:50
Classification of Meat-grinder Chucks using Image Processing

ABSTRACT. The work performed includes an end-to-end solution to a product classification problem with machine vision. Meat-grinder chucks separated by manpower within the company cause loss of time and errors. To solve the problem, chucks are detected with an infrared sensor, and the photos are taken. First the outer diameter of the chuck is calculated by Hough transform, after that, focusing inside the chuck, the radii of the inner circles are found. The circles found (in order to evaluate more accurately) are trained via a machine learning algorithm Naive Bayes method, the model is obtained. It has been determined in the study that this problem can be solved by a simpler method, mode operation. The output of the trained model indicates which class the product belongs to. The pneumatic pistons in the system are triggered by the time cutter software running on it and the products are separated to determined box. As a result of the tests performed on the system with the method of taking mode, it was determined that the model works with 99.2% accuracy. In the tests performed with the Naive Bayes method, an accuracy of 95.8% was achieved.

11:10
Development of Aluminum Conductor Design and Manufacturing Processes for Industry 4.0
PRESENTER: Erhan Sayın

ABSTRACT. Our industry, trying to meet market needs in a fast, flexible and efficient manner, is entering Industry 4.0, a new industrial reform, thanks to rapidly developing technology opportunities. Industry 4.0, the vision of the future, opens the door to product development in all software hardware, communication of production and service processes, real-time information exchange of machines and products, autonomous control and optimization. In order to be competitive in the market, there are two main axes: making a difference and low cost. The issue of creating difference is now perceived as innovating in terms of all functions of the business. Low cost means efficient use of resources, improving efficiency and avoiding processes and activities that generate waste (waste, waste, scrap). When the smart factory is considered; For a production system to be competitive, the network of connected devices (benches, devices, tools) and sensors is expected to form a lean manufacturing ecosystem.

Industry 4.0 focuses on increasingly individualized customer requests and includes services that include the entire chain, starting from the idea stage, from product development and production order to the distribution and recycling of a product to the end user. The purpose of this project is to develop this concept called 'Industry 4.0' in this period when new needs, automation, machine-to-machine communication and high technology have become an indispensable element, and to create and integrate the infrastructure in a way that completely meets the needs of our business.

For the first time in aluminum conductor and cable manufacturing in our country, the Industry 4.0 approach will be integrated into manufacturing processes. The algorithms and decision support mechanisms that will be created with the original contributions of our company constitute the innovative and original quality of the company.

11:30
Visual Feeding of Magnetic Cores for Winding Automation
PRESENTER: Ayberk Çelik

ABSTRACT. The influence of robot integration into production process is rapidly increasing. Feeding systems of intermediate products are one of the most important steps in ensuring full system automation towards fourth industrial revolution. Feeding systems generally consist of mechanical solutions. However, some production components cannot be fed to production cell with known methods due to mechanical restrictions such as fragility or sensitivity.

In this paper, novel background estimation and edge detection methods are proposed to provide reliable magnetic core feeding by visual inspection methods. Mechanical solutions were inadequate due to the irregular shapes of magnetic cores. Also, contortion of magnetic cores causes additional drawbacks. To overcome this problem, a method is developed to determine which products can be caught by the gripper and to check conformity of their shapes to increase production yield. The proposed method provides a decisive feeding process and ensures 100% accuracy.

11:50
PCB Solder Pad Defect Detection Using GLCM Features
PRESENTER: Sinan Özen

ABSTRACT. The influence of smart algorithms is rapidly increasing and getting more involved in industrial applications every day. To survive in the current competitive environment, short-production and short-lead times are important factors for any company in the electronics industry. As defects reduce the yield rate and increase the production costs, they need to be eliminated or fixed before the production process. In this paper, a machine learning algorithm based on combination of Gray Level Co-occurrence Matrix (GLCM) and conventional image features for PCB defect detection are presented and discussed. The proposed system is applied on the 1282 images whose are collected from manufactured products. Extracted features were classified using supervised Support Vector Machine (SVM) algorithm. The proposed method can classify defects with an average of 98% accuracy which is an acceptable percentage for industrial applications.

12:10
Visual Analysis of Turkey's Textile Sector with Gephi Complex Network
PRESENTER: Taner Ersöz

ABSTRACT. The Turkish textile and ready-made industry is mostly export-oriented and has an important place in our country's economy and the world market. In this study, according to the lower area of the textile sector in Turkey, it is aimed to create network analysis of export figures by network analysis of their productions according to region and provinces. Gephi 0.9.2 program was used in the study. Gephi is open-source software that is based on data mining and visualizes the complex relationships between data through a network structure. Sector production data were used from TOBB and export data was obtained from TURKSTAT. The weighted values of the producer information belonging to the regions according to the NACE classification of the textile sub-sectors were used on the Gephi program. It has been observed that the textile sector has spread to all regions, and it has been found that the regions where the sector concentration is highest are the Marmara, Aegean, and Southeastern Anatolia regions.

12:30-14:00Break
14:00-15:40 Session 7A: Parallel Session
14:00
Examination and Regulation of Market-Retail Warehouse with CRAFT and Entropy Method
PRESENTER: Emre Mutlu

ABSTRACT. The market diversifies the entire product range of supermarkets and offers them to the customers as door-to-door service. In the establishment of the Izmit warehouse of the brand, technical methods were not used when settling in the warehouse, which was rushed for some reasons. We used Craft and Entropy methods to improve the settlement. Thanks to the Craft method, we determined the locations of the aisles. Thanks to the entropy method, we determined the order of importance of the criteria of the aisles. The result is clear and clear, as the locations determined by Craft are based on flow and cost matrices. However, when we examine the system with Entropy, it is seen that the importance of the criteria is strongly determinant. The obtained Entropy method results helped the warehouse manager to find more detailed and more successful options for settlement.

As a result of our work, the flow created by warehouse employees between aisles was both arranged according to weighted combinations of orders and shortened. In this way, order preparation time was shortened and a faster transition to the transportation phase was achieved. Since the preparation time was shortened in the total time, the time to reach the user for the order was shortened, thus increasing the prestige of both new customers and the company. In this way, if a new change is required or if some of the aisles are to be relocated for other reasons, new ideas can be obtained by looking at the order of importance of the criteria. When the order is to be maintained, improvements can be made in the criteria properties of the aisles, again by looking at the foremost degree of the criteria. As a result, this work both optimizes the warehouse to current conditions and keeps it prepared for future changes. In this way, it increases the profit of the company, the employees are less tired and more motivated, and the formation of a happy customer base that receives their orders quickly is served.

14:20
Examining the Relationship between Unlimited Improvement and Service Quality: A Private Sector Study in Libya ( Sınırsız İyileşme ile Hizmet Kalitesi İlişkisinin İncelenmesi: Libya'da Bir Özel Sektör Araştırması)
PRESENTER: Orhan Küçük

ABSTRACT. The concepts of service quality and unlimited improvement are key tools that are used in business administration to measure the quality of the service provided to the internal and external customers. Moreover, the relationship between these concepts can be decisive for the performance of the organization. The concept of unlimited improvement is a quality management framework developed by Küçük (2011). The concept is based on the dimensions of total quality management and uses similar techniques of implementation (Küçük & Küçük, 2012). The aim of this research is determine to relationship between service quality (Parasuraman-ServQual) and unlimited improvement (UI) within the engineering firms in Libya. In this study, the data will be compiled using Face-to-Face questionnaire method. Research will be conducted on companies to be selected by the random sampling method among the total oil companies willing to participate in the study. By testing the validity and reliability of the obtained data, the relationship between unlimited improvement and service quality will be determined. SPSS 23.0 package program will be used for this. As a result of this study, the relationship of unlimited improvement practices with service quality will be determined, thus the importance of the implementation will be revealed. Suggestions will be shared to those concerned in line with the information obtained.

Sınırsız İyileşme ile Hizmet Kalitesi İlişkisinin İncelenmesi: Libya'da Bir Özel Sektör Araştırması Hizmet kalitesi ve sınırsız iyileştirme kavramları, iç ve dış müşterilere sağlanan hizmetin kalitesini ölçmek için işletme yönetiminde kullanılan temel araçlardır. Dahası, bu kavramlar arasındaki ilişki organizasyonun performansı konusunda belirleyici olabilmektedir. Sınırsız iyileştirme kavramı, Küçük (2011) tarafından geliştirilen bir kalite yönetimi çerçevesidir. Kavram, toplam kalite yönetiminin boyutlarına dayanır ve benzer uygulama tekniklerini kullanmaktadır (Küçük ve Küçük, 2012). Bu araştırmanın amacı, Libya'daki mühendislik firmalarında hizmet kalitesi (Parasuraman-ServQual) ile sınırsız iyileştirme (UI) arasındaki ilişkiyi belirlemektir. Bu çalışmada veriler Yüz Yüze anket yöntemi kullanılarak derlenecektir. Araştırmaya katılmak isteyen toplam petrol şirketleri arasından rastgele örnekleme yöntemi ile seçilecek firmalar üzerinde araştırma yapılacaktır. Elde edilen verilerin geçerlilik ve güvenilirliği test edilerek sınırsız iyileştirme ile hizmet kalitesi arasındaki ilişki belirlenecektir. Bunun için SPSS 23.0 paket programı kullanılacaktır. Bu çalışma sonucunda sınırsız iyileştirme uygulamalarının hizmet kalitesiyle ilişkisi belirlenecek ve böylece uygulamanın önemi ortaya çıkacaktır. Alınan bilgiler doğrultusunda öneriler ilgililerle paylaşılacaktır.

14:40
THE GRANGER CAUSALITY ANALYSIS OF CONSTRUCTION INDUSTRY: CASE STUDY IN TURKEY

ABSTRACT. The general progress of economy is being associated with the main industries of countries. Different kinds of them have been taken the lead, in the worldwide. Furthermore, the situation may depend on the structure of the countries in here. Turkey possesses many opportunities in obtaining materials from different resources and hot money flow. However, this type of cash flow and easy access to resources cannot be seen all the time. Therefore, precautions have to be taken. Industries must come to a level that they will not always rely on these types of flows, but which one of them have to be the locomotive? In Turkey, the construction industry is considered to have taken this position. The purpose of this paper is observing this hypothesis to understand the relationships of other main sectors and construction industry. Granger causality tests have been performed for 19 main sectors and the industries having relationship are investigated in detail that how much relation is observed between them

15:00
WHAT EFFECTS CRYPTOMONEY EXCHANGE?

ABSTRACT. The crypto money exchange, which differs from the known stock exchanges but has similar sides, is a platform that can be bought and sold crypto money. There is a different working system than other stock exchanges, especially because of the absence of an intermediary institution. Along with this, different kinds of views has been put forward in effects on crypto money exchange. Derived from this reviews, this study has been emerged. The main purpose of this study is to find out which factors are affected by the volume of crypto money exchanges and to discover which factors should be considered when a choice is made in this direction. Visitors, Pairs, BTC, Twitter Followers, Liquidity, Scale, Withdrawal, Coin Number, ETH, Fees, Alexa Order, API Coverage and Regularity Comp. has been selected as candidate effects. Multiple Regression has been utilized for understanding the relationship with the scale of Cryptomoney exchange. As a result, the course of the currency directly related with the topic of Visitors, BTC and Liquidity. According to the situation of the real life, it is clear that manipulators and big investors will play too much about the general trend of the stock exchanges.

15:20
An Application Scenario of Blockchain Technology in Nuclear Security
PRESENTER: Pelin Üstünyer

ABSTRACT. Nuclear security is to take all necessary physical and digital protection measures to prevent, detect and respond to theft, sabotage, and other malicious acts that may be carried out from inside and outside during the use, storage, and transportation of nuclear materials. In this study, it is aimed to present a proposal that will provide nuclear security. With this purpose, blockchain technology has presented as a database structure that provides security, privacy, and authentication to prevent cyber-attacks and digital privacy violations in the nuclear sector. In the study examining blockchain types, Federated (Consortium) Blockchain determined as the most suitable blockchain type that can be applied to the nuclear security area. Besides these, although nuclear security and nuclear safety have different focal points, they should be handled together because they cover each other. The work to be done for any of them should not affect the other one negatively. For this reason, this study has studied from a systems engineering perspective.

14:00-15:40 Session 7B: Parallel Session
Chair:
14:00
Automated Inspection Approach For Remanufacturing

ABSTRACT. Inspection in remanufacturing is a labour-intensive and time-consuming step that involves identifying defects on the surfaces of components that are candidates for remanufacturing. Traditional techniques have limitations in terms of cost, detecting multiple defects at a time, and inspector reliability. As a result, automated inspection techniques have garnered remanufacturers’ attention because of their potential cost advantage and improved defect detection capability. This study examines the capability of optical inspection techniques to decrease inspection costs and errors in remanufacturing. We implemented object detection methods to classify and locate defects on steel surfaces from surface images with reasonable accuracy. The YOLO (You Only Look Once) V4 algorithm was used to capture and classify the defects. The performance of the algorithm is compared with state-of-the-art approaches using recall, average precision, and mean average precision metrics. Our model demonstrates effective defect localization with a mean average precision (mAP) of 64.13%, which shows promise for the development of automated inspection technology.

14:20
Assessment of Factors Affecting Dam Occupancy Rate Using Fuzzy Cognitive Map with Extended Great Deluge Algorithm

ABSTRACT. Water scarcity is one of the most crucial challenges the world has ever faced. Protecting potable water resources and meeting the water needs of the society has become a vital issue discussed in recent years. In order to overcome this challenge, it is important understand the factors affecting water resources. In this study, some of the factors affecting the occupancy rate of Istanbul dams are examined using fuzzy cognitive map (FCM). The FCM are trained via extended great deluge algorithm (EGDA). This paper mainly aims 1) to test the performance of FCM methodology on the relevant topic and 2) to reveal the casual relations between studied factors. The results of the study can indicate how FCM can represent casual relationships for the relevant case and also help to increase our understandings on the factors affecting occupancy rate of dams.

14:40
Development of an Ontology for Defect Classification in Remanufacturing

ABSTRACT. Remanufacturing within the automotive industry has become an important part of environmental and reusability efforts. A primary function of remanufacturing is the correct identification and classification of defects on the product. This paper will look specifically at cylinder heads within remanufacturing applications to validate and summarize the defects most commonly found on the respective components. A thorough ontology was developed based on a literature review to define, distinguish, and prioritize various defects for remanufacturing. Furthermore, expert opinions within the industry and text-mining were sought to confirm and extend the ontology for defect classification in remanufacturing. The results show that these methods provide sufficient supplemental knowledge to validate critical or underdeveloped areas of the ontology. Expert opinions were extremely valuable in communicating information that is not discussed in scholarly articles and validating information found in scholarly articles.

15:00
ANALYSIS OF FACILITY LAYOUT DESIGN PROBLEM IN DEFENSE INDUSTRY COMPANY

ABSTRACT. In today's competitive conditions, businesses in the manufacturing or service sector try to reduce their costs as much as possible in order to survive under these competitive conditions. Businesses make efforts to reduce costs at many stages, from determining the most suitable factory location, to minimizing stocks, from the production system that prevents waste as much as possible, to the most appropriate placement of machines or work areas, to logistics activities. It is aimed to increase productivity and efficiency while reducing costs. Studies carried out so far show that the extra flow of materials, information and documents made between machines or work areas in an enterprise causes unnecessary transportation, thus increasing costs. These unnecessary transports are usually due to the machines or working areas not being properly placed within the business. If the business places the machinery or workspaces properly, it can reduce production costs. This can also increase the strength of the business in the market.

If we explain the facility placement scheme; Manpower can be defined as planning and integrating the flow paths that involve the production of the product or its parts, passing through the production stages and sending them as products in order to provide the most efficient and economical relationship between machine and material. According to the above definition, we can define in-plant arrangement as finding the physical relationship between machinery, manpower and working areas and arranging the production elements in the most appropriate and efficient way for the business purposes and product qualities.

15:20
Implementation of Sales Demand Forecasting in An Automotive Company
PRESENTER: Onur Canpolat

ABSTRACT. The primary purpose of the existence of companies is to ensure continuity, to satisfy their employees and customers, as well as to make a profit. The most effective way to achieve customer satisfaction is to send the goods and services requested by the customer on time. To send the orders at the desired time, it is necessary to estimate how many orders will be in which periods. Because of the many factors that affect the sales demand, accurate and reliable estimates are required. Dollar rate, GDP, number of car parks, number of vehicles produced, number of exports, interest rate, CPI, and PPI are taken as factors affecting the engine sales demand. In this study, the sales volume of the engine bearing, one of the automotive sub-industry products, was calculated using the multiple regression model and several methods in time series analysis and these methods were compared with each other.

15:40
LNG Supplier Country Selection with Fuzzy TOPSIS Software
PRESENTER: Onur Canpolat

ABSTRACT. Energy is one of Turkey's most important import item and gas imports is also a pioneer in this regard. Liquefied Natural Gas (LNG) is a special odorless and colorless natural gas form. LNG stands out as an important alternative in the transfer of natural gas between countries. Due to providing alternative opportunities, LNG is seen as an increasingly important source of energy these days for countries that import large quantities of natural gas, such as Turkey. However, the continuous variability of economic and social relations between countries makes determining the country where LNG will be imported into a critical and important decision. This issue is a strategic decision for countries. Therefore, an accurate assessment should be made under the current conditions and the most suitable supplier country should be selected. In this study, criteria and alternatives were determined by examining the reports of many different institutions such as the Energy Market Regulatory Authority (EMRA), the Ministry of Energy and BOTAS. Fuzzy AHP was used in the evaluation of criteria and alternatives, and Fuzzy TOPSIS method was used in determining the most suitable supplier country.

15:40-15:50Break
15:50-17:30 Session 8A: Parallel Session
15:50
Analysis for Dominant Air Carriers of U.S: A Look to Future

ABSTRACT. Financial and administrative activities are important internal factors of company policies in the field of transportation. Land transport, air transport and maritime transport, which are transport area, have their own administrative and financial characteristics. In this study, the data published by the top five airlines operating in the United States of America is used to provide an insight into the managerial policies that companies are currently implementing. In line with the gather data, managerial policies of companies are evaluated with AHP and TOPSIS method, which are the most widely used multi-criteria decision making methods. The managerial policies are evaluated and categorized accordingly and the deficiencies and good aspects of the company policies are provided.

16:10
Location Based Turkish Natural Language Study, VOICE OF SAKARYA
PRESENTER: Furkan Saracoglu

ABSTRACT. The aim of this study is to make the thoughts, feelings and emotions of the people who make up the mining region, environment and society understandable by analyzing the textual data that is constantly updated and rapidly growing data world. Within the scope of the project, the tweets obtained through Twitter, which has an important place among social media tools, were predicted using natural language processing and machine learning algorithms and methods. This study, which will lead to listening to the natural voice of some of the people by aiming less cost, fast access and results instead of various survey studies. Project has been supported by different software and different approaches. It was created to automate the steps of the study so that certain results and gains can be achieved by reaching a conclusion in the long term, not in only the short term. In this study, in which various software and applications were used, the performance was kept at optimum and the least data loss and error rate were studied. Database systems are used for data protection and backup operations. Data that has been cleaned and transformed into information is transformed into a meaningful visual with data visualization application and solutions and presented as a monthly visual report. For ease of access, it is aimed to ensure that everyone can access analysis results by using cloud and web systems.

16:30
Benchmarking Study of Firm-Level Innovation Capability Assessment Tools
PRESENTER: Seçkin Dilek

ABSTRACT. Academic literature and non-academic (grey) literature frequently address that organizations should improve their innovation capability, but it offers limited guidance on how it can be assessed and improved. Although a significant number of maturity/assessment models have been proposed for assessing and improving an organization’s innovation capability on the firm-level, the research and practice on this topic is still in its emergence, characterized by a scarcity of empirical validation. In the last decade, researchers and practitioners have proposed assessment/maturity tools, models or frameworks to assess and improve organizations’ innovative capabilities but firms in different sectors and sizes are not able to know which innovation assessment tool has the best fit for their organizational structure. Assessing innovation capability by using the popular and generally accepted innovation assessment tools is inadequate for managing innovation. Some innovation assessment tools deliver the detailed output report automatically about the innovation capability of firms. However, some tools deliver only spider diagram with different dimensions of innovation capability in firm-level.

In this study, the most popular 10 innovation assessment tools developed in different countries (including Turkey) are presented in detail. In addition, three of these innovation assessment tools (improve, inno-survey, USIMP Innovation Report Card) are benchmarked systematically using six specific metrics (such as dimensions, result systematics). We expect this benchmarking to be useful for institutions and firms (especially SMEs) while there is absence of an explanatory guide to choose the most suitable innovation assessment tool that fit for their organizational structure.

16:50
Call Center Performance Evaluation using K-Means and Self-Organizing Map Clustering Algorithms
PRESENTER: Nevra Akbilek

ABSTRACT. Call center agents are trained on how to respond or behave in projects where both types of calls are made (inbound call, e.g. support services and outbound call, e.g. membership campaigns). However, situations that cannot be planned before in the process may cause unnecessary extension of call times. The inefficient time of the call center staff can depend on two reasons: process problems or unwanted staff activities. On the other hand, some personnel activities which exploit the system have a negative impact on productivity. This study aims to detect and make sense of anomalies in the processes and to take necessary actions about these anomalies using the k-means and Self-Organizing Map (SOM) clustering algorithms. The results of the research revealed the basic characteristics of the anomalies about agents and team leader activities that occurred in the call center. Also, required actions to eliminate these characteristics are proposed.

17:10
Active Control of the Hyperchaotic Supply Chain System

ABSTRACT. A supply chain dynamic system exhibits extremely complex, so-called hyperchaotic, behaviors, with the immediate, unpredictable or undetermined effects of state variables such as demand, production quantity and stock, initial conditions of the system or parameters. In real life, these hyperchaotic behaviors for a supply chain system reduce competitive power in the market, increase costs, and may even cause the system to collapse in a very short time. Therefore, specific control methods for retaining steady state of the supply chain dynamic system have been developed. In this study, active control method was applied to a hyperchaotic supply chain system. Mathematical application steps of active control method and simulation results are given in the relevant sections.

15:50-17:30 Session 8B: Parallel Session
15:50
Kernel Density Estimation for Tracking Important Topics on a Map during the Earthquake

ABSTRACT. Social media platforms have gained importance for the potential usage of location information that can be retrieved from these platforms. In addition, analyses of data retrieved from Twitter become a crucial task in order to have better understanding of specific topics. Therefore, Nepal earthquake data-set is taken into consideration due to understand general situation of society during and after earthquake disaster. In the proposed study, topics are demonstrated in relation with locations. In order to accomplish aforementioned aim, latent dirichlet allocation (LDA) is applied on location specific tweets. After that, for the hotspots identification through important topics of each location, kernel density estimation (KDE) is utilized. For the visualization on map, Basemap Python library is implemented. All in all, important topics are extracted and demonstrated on a map in order to provide representation of the situation for the earthquake.

16:10
Innovation Management using Spherical Fuzzy Set-based MADM: A Case in the Turkish Banking Sector
PRESENTER: Senay Demirel

ABSTRACT. Innovation is an important instrument for actors in financial sectors, as in many other sectors, to compete in challenging conditions and provide innovative solutions to their clients. They realize many innovation projects, especially for reasons such as improving the products of banking and increasing their financial strength. Since the open innovation method introduces a framework for gathering many different project ideas, there is a need to select the most suitable ones in the context of innovation management. The department supervising the innovation management has limited resources that can be dedicated to this effort so that decision analysts must evaluate various project ideas and determine a proper ranking. In case many attributes should be considered in assessing the alternative innovative ideas, they can be efficiently and effectively ranked via multiple attribute decision-making (MADM) methods. When the decision process needs several experts, the decision process can be very complicated and renamed as Group MADM (G-MADM). In the project idea selection problem, the alternative ideas are required to be evaluated by the decision-makers. To model the uncertainty and vagueness in their judgments, fuzzy-based G-MADM methods are developed as beneficial tools because of their representation power in quantification. A recent fuzzy set concept introduced in the literature is spherical fuzzy sets (SFS) which can simultaneously model the positive and negative opinions as well as the possible hesitancy of the decision-makers. For considering the extensive human judgment representation power of SFSs, Analytic Hierarchy Process (AHP) and Weighted Aggregated Sum Product Assessment (WASPAS) tools are utilized under spherical fuzzy environment. SF-AHP obtains the importance of the attributes which are currently used in real-life innovation management applications maintained in the relevant committees of the bank. Ten alternative innovation ideas are evaluated by SF-WASPAS to select the most appropriate one for the bank and its customers.

16:30
Airline passenger planes arrival and departure plan optimization using evolutionary strategy
PRESENTER: Suraka Dervis

ABSTRACT. Air travel became more attractive since humanity has learned to fly, and it is preferred for long distance travels because long distances are covered in a short time. Air travels continuously increased in Turkey and the world. The volume of passengers in Turkey increased in domestic and international lines. Synchronization and optimization of international and domestic flights are important for passenger maximization. Some of passengers don’t stay in the city of the main Airport in the country or there is no direct flight to their destination, so they have to book two flights at least. But naturally humans don’t like waiting. In case of long transit time they tend to book from another airline company. In this study, we will try to solve this problem by using the evolutionary strategy to maximize the number of total passengers saved between arrival and departure. In our problem we have arrival and departure plan. Mutation will be applied to the gens selected randomly, then the selected gens will be shuffled and re-placed.

16:50
A Machine Learning Model Proposal for COVID-19 Patients' Diagnosis and Survival Estimation

ABSTRACT. COVID-19 caused by SARS-COV-2 virus was first in China's Wuhan Province and has caused the pandemia by spreading all over the world. Machinery learning and artificial intelligence practices also increased folding with Covid19 pamdemic. In this study, COVID 19 diagnostic and survival estimates have been planned to develop the patients with clinical and demographic data in our hands with clinical and demographic data and processing clinical data with clinical data and providing the right and right diagnostic with clinical data by processing clinical data in the next period. First of all, a data set showing the characteristics and positive outlining situation that provides or providing COVID19 positivity. In this data set, the person's contact query, detailed epidemiological properties, detailed clinical data, hospitalization and treatment processes or foot-tracks, laboratory tests are determined, after the Excel table was made of the system. In particular, the number of cases is kept as wide as possible, diversification can be made depending on the new features. Then the meaningful features will be re-examined with the machine to retract from the data set. With a good system learning, the percentage of the people who are considered to be a sign of disease, the percentage of being patient will be calculated and the survival will be provided. A threshold value is determined on this possibility, it will be recommended to test the threshold value. Thus, the survival for epidemiological characteristics / clinical findings and transmission risk factors will be removed by the combination of combination. In order to make the data set ready for machine learning, the priority of the lacking and contrary data between the data will be performed and the property reduction studies will be carried out. After the data precipitation step, the method of support according to the feature status of the data, the method of vector machinery, decision trees or artificial nerve networks are tried to be determined. The best model will be predicted with future data.

17:10
Robust Ship Heading Control using Hyperbolic Tangent Function-based Adaptive Sliding Mode Controller

ABSTRACT. Due to the uncertain and nonlinear effects of ship maneuvering, ship turning motion is a key topic in autopilot systems to provide the safety of ship sailing in an open sea. Therefore, this paper presents the robust heading control design of a nonlinear ship model using the hyperbolic tangent function-based adaptive sliding mode control (SMC) structure. In the designed scheme, the adaptive SMC control law is derived to increase the robustness of the control performance. The hyperbolic tangent function is also used instead of the switching function to guarantee rapid convergence and eliminates the chattering problem. The simulations are realized in MATLAB environment with step and sinusoidal reference heading angles. The obtained results show that the adaptive heading control method provides robust and efficient control performance for the different reference trajectories.

17:30-17:40Break
17:40-19:20 Session 9A: Parallel Session
17:40
Determination of Coding and Non-Coding RNAs Using AVL-Tree Based Protein Mapping Method and Deep Learning

ABSTRACT. Coding and non-coding RNAs play important roles in various cellular activities, diseases and analysis of new transcriptions. With the development of next-generation sequencing technology, coding and non-coding RNAs can be distinguished quickly and accurately. However, these developed methods cannot analyse small-sized coding and non-coding RNAs very well. Because of this problem, the effect of computational-based approaches in this area has started to increase. In this study, by applying computational-based approaches, the distinction between human-coding and non-coding RNAs was carried out with a deep learning model. The study consists of four stages: obtaining coding and non-coding RNA data, mapping the data with protein mapping techniques, applying the deep learning model, and determining the performance of protein mapping techniques with evaluation criteria. In the study, EIIP, hydrophobicity, integer, CPNR and AVL-based protein mapping techniques were used to convert proteins into numerical expressions. At the end of the study, all protein mapping techniques performed a successful classification process. With the proposed study, it has been observed that the selected protein mapping techniques and deep learning model are effective in the separation of coding and non-coding RNAs.

18:00
Customer Relationship Management using Deep Learning in Workspace Environment
PRESENTER: Nevra Akbilek

ABSTRACT. In this study, CRM data of an e-commerce enterprise are analyzed. The dataset that contains customer complaints and demand entries is used for sentiment analysis via an RNN deep learning model. In this context, customer opinion is predicted based on the Gated Recurrent Units (GRU) neural network (NN) method. For this purpose, different models were created using the combinations of optimization method (Adam, Rmsprop, etc.), activation function (sigmoid, ReLU, etc.), and neuron numbers. Different models are tried, and the best model for sentiment analysis was revealed by comparing the performances of the model combinations created. After the entries taken from the system are passed through various cleaning steps, customer opinions are classified as positive and negative classes with an approximately 0.9 accuracy rate is obtained.

18:20
The Fuzzy Logic Approach to Determine Education Method in Organizations
PRESENTER: Buşra TaŞkan

ABSTRACT. Education is a phenomenon that causes behavioral changes in the person given. Education, which is a tool used by organizations to develop their employees and hence themselves, is also one of the primary methods used to overcome crises, pandemics and disasters. In addition, in-service trainings are also used as a motivation tool for employees. In-service trainings cause loss of time for organizations despite their mentioned benefits. Therefore, it is necessary to optimize the education method, duration and cost. Based on the mentioned importance of the subject, in this study it is aimed to optimize the decision problem related to that how in-service trainings should be given in organizations (face-to-face education or distance education). The fuzzy logic was used as the method because it is suitable for the nature of the problem under consideration. For this purpose, a fuzzy system with 5 inputs and 2 outputs was created and solved with MATLAB's fuzzy logic toolbox.

18:40
Risk Management for Disasters Using Fuzzy FMEA Method
PRESENTER: Buşra TaŞkan

ABSTRACT. The continuous development of technology has brought a different dimension to the ongoing disasters throughout history of humanity. Disasters are no longer just natural-induced, man-made disasters are also happening. The Covid-19 epidemic we are experiencing nowadays is also an obvious example of this situation. Due to the current importance of the subject, the risks related to disasters have been evaluated in this study. For this purpose, risks related to disasters were analyzed with key success factors (Technological, Economic, Operational, Social, Legal, Environmental, Institutional) in the disaster management cycle using the Fuzzy FMEA method. In this study, it is aimed to provide a perspective in order to determine the strategies for reducing the risks of disasters.

19:00
An Intelligent Psychiatric Recommendation System for Detecting Mental Disorders
PRESENTER: Esma Nur Ucar

ABSTRACT. Despite the remarkable recent developments in mental healthcare, many uncertainties in the diagnosis process remain. Even a detailed, well-timed, and closely followed psychiatric interview may not be sufficient to produce an accurate differential diagnosis. At the same time, an insufficient number of specialists and the resultant heavy workloads impede diagnostic efforts, making it very difficult to receive appropriate medical services and manage the treatment process. Such problems underscore the need for auxiliary systems to help experts in making diagnoses, saving both labor and time. For this reason, we propose a new intelligent psychiatric recommendation system with the Comprehensive Psychiatric Differential Diagnosis Test (CPDDT), which we created to screen and differentiate among psychiatric diagnoses. To guide expert in using the system, we included axis one and axis two diagnosis groups, which respectively refer to clinical and personality disorders in the DSM-4. The goal was to measure areas affecting the course of an illness and the treatment plan developed by a specialist, including functionality, memory, and suicidal thoughts. The CPDDT can detect 48 different diagnostic groups from the answers to 319 questions. The system was subjected to an online test of 676 users via a web system developed by DNB Analytics. Psychiatrists evaluated the results in a clinical setting. The test results were then evaluated by the evolutionary simulation annealing LASSO logistic regression model. After determining the importance of each question on the scale, the algorithm eliminated the questions with the least impact and the test was reduced to 147 questions, producing a .93 level of accuracy. In addition, the algorithm found the probability of each patient suffering from a disorder. In summary, the new machine learning-based CPDDT was finalized to include 147 questions; the algorithm is presented here as a useful suggestion system for experts engaging in the diagnostic process.

19:20
Characterizing Risk Factors of 2019 Coronavirus Disease (COVID-19) Using Prescriptive Analytics Techniques
PRESENTER: Yeliz Çotoy

ABSTRACT. Since the first case of the pneumonia caused by 2019 novel coronavirus (COVID-19) is found in Wuhan, studies conducted on patients with COVID-19 disease in the literature were reviewed all over the world. Common characteristics of patients with COVID-19 were clarified. Demographic characteristics, coronavirus disease 2019 symptoms, laboratory evaluations, and clinical management were abstracted. There was no geographical limitation or any other restriction in this study and these risk factors were divided into two main parts as external and internal factors. These factors were categorized; External factors such as indoor and outdoor air pollution factor and COVID-19 factor of that region (Case fatality ratio and recovery rate) and internal factors such as preexisting diseases, pregnancy, demographic characteristics like older age. This study aims to quantify the associations of COVID-19 and characterize epidemiology and risk factors to make predictions before contracting this disease. However, when prescriptive analytics is used to make predictions in addition to statistical analysis, a tool that helps health systems and physicians is developed.

17:40-19:40 Session 9B: Parallel Session
17:40
Bottleneck Station Scheduling using Ant Colony Algorithm in a Tire Factory

ABSTRACT. Correct production planning is critical for surviving in a competitive environment and meeting customer expectations on time. Planning can become much more complicated in sectors where there are many products, such as tire production. This study focuses on a machine that causes a bottleneck in the production of a tire factory. With this machine called Quadruplex Extruder, rubber is extruded and transformed into a Tread material product mainly responsible for some of the essential tire features like low rolling resistance and brake distance. This study aims to minimize the setup times in the production by optimizing the manufacturing order of the products to be produced in this machine, with the Ant Colony Algorithm (ACO), a metaheuristic method. The randomly generated schedule was compared with the schedules produced by the ACO. As a result, it has been shown that the ACO can provide fast and suitable solutions to decision-makers in production planning.

18:00
Examination of the causes of traffic accidents using decision trees: A Case Study in Sakarya Province in Turkey

ABSTRACT. Although recent improvements have been achieved in terms of the mortality and injury rates due to traffic accidents in Turkey, traffic and road safety stand as a problem that continues to affect social life quite deeply. In this study, traffic accidents occurring in Sakarya Province in Turkey were examined. The data for this study were obtained from Sakarya Provincial Police Department. There are 139 feature classes related to the accident that took place in the data set. In this study, firstly, the important features within data set were revealed to classify the data set. Secondly, classification algorithm frequently used in the literature to achieve classification goals; decision trees, were utilized to discover more accurate information and relationships from the data set. Performance evaluation of these algorithms was carried out according to classification quality indicated by the results.

18:20
Interaction of artificial intelligence with industry 4.0 components in manufacturing sector

ABSTRACT. Industry 4.0 is an intelligent system used as a flexible production line for almost all production processes of real-time information provided by Artificial intelligence (AI), Internet of Things (IoT) and other digital technologies. Artificial intelligence sharpens business intelligence, an important development for the global economy. The purpose of this paper is to show how industry 4.0 components interact with artificial intelligence and examine how industry 4.0 components affect artificial intelligence. For this purpose, a questionnaire containing industry 4.0 components has been developed and this survey has been applied to companies operating in the manufacturing sector. The obtained data were analyzed with the spss 22.0 statistics software and the results were presented. It has been determined that the Internet of objects, smart systems and technologies, the digital factory which are among the components of Industry 4.0, are closely related to artificial intelligence.

18:40
IMPROVING QUALITY CONTROL PROCESS WITH FMEA ANALYSIS IN THE AUTOMOTIVE INDUSTRY
PRESENTER: Safiye Turgay

ABSTRACT. Competition, which is a result of rapid development and globalization in the light of technology and science, makes it difficult for companies to survive. Companies are looking for new solutions. For this reason, businesses have been in search of different systems and solutions. This is why the Total Quality Management is adopted in the management approach. Today's management principles have led to the adoption of Total Quality Management. Total Quality Management (TQM) is an effort to continuously improve the quality of goods and services with the participation of all employees, based on customer satisfaction and in order to meet customers' demands. It is based on customer satisfaction and meeting customer expectations.

It is very important for businesses to be able to make both new customer acquisition policies and repurchase action to the retained customer. Research has shown that re-purchasing of existing customers has a much higher return than acquiring new customers. The maximum level of customer satisfaction will prevent existing customers from turning to other companies. Customer complaints should be properly managed for customer satisfaction. '' According to research, approximately 96% of dissatisfied customers do not complain and most of them try other ways until they leave the business (Barış, 2006). is of greater importance. Complaint; It is a feedback for the continuous improvement of the products foreseen with a total quality understanding, a tool for increasing the performance of the product and an opportunity for the business (Alabay, 2012).

Our study aims to increase the detection of errors in the production process at the eye control point in order to turn the customer complaints of a company that performs 100% inspection into an advantage.

In the application part of the study, incoming customer complaints are classified according to their types and effects, and the effects of error types are analyzed. Casting errors with the highest priority number were emphasized. Suggestions are presented to minimize casting errors. Considering that hundreds of thousands of companies are active and competitive in the globalizing world, it becomes very difficult for companies to maintain their continuity. For this reason, companies should accelerate their growth and development activities, taking into account their business potential and customer policies. Customer satisfaction is the most important factor for the growth and development of businesses. Complaints made by customers justify the solution of the problems we are working on. In the theoretical part of our study, by referring to the importance of visual inspection and quality control and determining the problems in the quality control process of the factory, alternative solutions were determined in the visual inspection process.IMPROVING QUALITY CONTROL PROCESS WITH FMEA ANALYSIS IN THE AUTOMOTIVE INDUSTRY

19:00
A Framework for Planning Logistical Alternatives in Value Stream Design with Milk-Run Approach
PRESENTER: Safiye Turgay

ABSTRACT. The manufacturer sends its orders to a specific logistics provider and leaves the whole process until the delivery to the customer to the planning of the subcontractor. Notifications to suppliers and logistics service providers are usually made one day in advance, and the tours to be made vary according to this information.

In the Milk-Run concept, tours are planned on a cyclical basis. Before a product is delivered to a customer or a consolidation center, it passes through multiple suppliers. The underlying idea of ​​this system of layout and repetition is to reduce the variability of processes and shipping costs. Another advantage of the Milk-Run concept is that it allows the transfer of products with less quantity than with Point-to-Point Transfer with more frequent cycles. Compared to Field Routing Services, transfer times are shorter and the process until the delivery to the customer is more transparent, which affects the quality positively.

The sample problem on hand describes the change in the current distribution and collection systems of a logistics service company. The response of variables such as duration, weight, volume and pallet was measured under various scenarios, together with huge cost and time savings by applying Milk-Run optimization. For the Milk-Run optimization application, an excel file extension named Log-Hub, and previews and outputs of the routes are presented. Based on this, 30% and 50% decreases were made on the constraints for six different scenarios, and the outputs resulting from these decreases and their comparisons with each other were given as the result obtained through the application.

As a result, suppliers may reduce their safety stocks, which in turn affects the fall in parts prices in the long run. With the effect of this situation, logistics service providers may reduce their transportation costs in the medium term by balancing the empty cycles they have to do.

19:20
GENETIC ALGORITHM-BASED APPROACH TO FLEXIBLE WORKSHOP PRODUCTION PROBLEM
PRESENTER: Safiye Turgay

ABSTRACT. Unlike the problems we have dealt with in the literature, it is inevitable that most of the problems encountered during production in daily life, the whole or a large part of the part is processed more than once in more than one machine. This situation not only provides success in complex production systems, but also requires the most efficient use of business resources.

The aim of the study offers solution methods, especially mathematical solution and genetic algorithm, by considering the preparation times, part movements, transportation times, quality control, maintenance and repair of other parts in the family of parts that the manufactured part is connected to, as flexible workshop type scheduling suitable for the workshop type manufacturing environment. For this reason, it is necessary to create a dynamic flexible workshop production environment, determine the appropriate standard times for each workpiece and make the most appropriate operation-machine assignment, and based on this, the appropriate production plan must be created. In this context, genetic algorithms are used for decisions to be taken, especially when making operation-machine assignments

19:30-20:00 Session 10: Keynotes
19:30
How to Use the Complex Networks on Machine Learning for Extracting Meaningful Patterns from Big Data

ABSTRACT. Artificial intelligence is everywhere, and it is crucial if we solve complex problems in big data. For example, Netflix needs to recommend a new movie for each personnel on time with high accuracy. Otherwise, everything is about theory and dreams in artificial intelligence. Machine Learning (ML) and Complex Networks (CN) could be used as effective intelligent tools to extract meaningful insights from big data to solve this problem. For more detail, ML researchers focus on individual observations, which do not consider the interactions, which ignore the functional roles of observations in the system. When CN researchers capture interactions in the networks, they give general structural patterns, and it is tough to use in the business. Therefore, we proposed the Complex Networked Operational Strategy (CoNOS) framework for solving different business problems. This framework proposes a novel approach that combines CN and ML for proposing new operational strategies in the business. In the framework, a new entity similarity function is proposed to estimate relationships/links among interactive observations. With this similarity function, we estimated 100 million relations in 17 seconds without using any parallel processing. Then, we proposed a new network embedding approach and shared an embedded dataset with machine learning approaches. As the results, we compared three approaches: (1) classic approach that uses traditional machine learning techniques; (2) the networked approach that uses machine-learning techniques on the network topology; and (3) the unified approach that combines network topology and real data with machine learning methods. In conclusion, we finally presented how the framework improves ML approaches by using CN. Companies can see hidden patterns in real-time and high accuracy and solve their business problems with high impact.