ABSTRACT. The traveling salesman problem (TSP) is a well-known combinatorial optimization problem with numerous practical applications. Various extensions, such as the general TSP, the vehicle routing problem, and the many-visits TSP (MVTSP), have also been studied. This paper focuses on the MVTSP, where the salesman may visit each city multiple times as required. We develop a heuristic method for solving the MVTSP and conduct computational experiments to compare it with existing approaches. Finally, we discuss several interesting properties of this problem.

ABSTRACT. With the increasing promotion of the concept of resource sharing, the Collaborative Vehicle Routing Problem (CVRP) has gained significant attention in recent years. Through resource sharing, it is possible to significantly reduce operational costs and improve logistics efficiency. Specifically, optimizing resource allocation, reducing empty miles, and minimizing transportation costs have become key factors for companies to achieve sustainable development and com- petitive advantage in modern logistics systems. This paper investigates a novel multi-depot open-path vehicle routing problem—referred to as the Shared Depot Multi-Objective Open Vehicle Routing Problem(SD-MDOVRP). This problem expands on traditional multi-depot vehicle routing problems (MDVRP). In this problem, goods are allocated within the path network, and vehicles depart from different depots, with the ability to return to any nearby depot, thus enhancing scheduling flexibility and transportation efficiency. To more accurately reflect the potential uncertainties in the transporta- tion process, this paper presents a Mixed Integer Linear Programming (MILP) model aimed at optimizing path selection, vehicle scheduling, and depot resource allocation, maximizing transportation efficiency while minimizing operational risks. Finally, experimental analysis is conducted to explore the relationships between the multiple objective functions, further investigating the trade-offs and optimization strategies among these objectives.

ABSTRACT. In this paper, we explore a heuristic algorithm to deal with the square packing problem constrained by no-overpass conditions, considering two types of horizontal partitions. The goal is to select the minimum width from a width candidate list while packing all given squares and satisfying the no-overpass constraints and partition board restrictions.

This paper proposes a constraint handling mechanism that prevents items from overpassing each other by optimizing item placement order within a given area. It then proposes a heuristic algorithm that first fills the bottom of the base material with row candidates generated by placing squares horizontally. Then, for each combination of such row candidates, it constructs slots to fill the remaining space efficiently.

The effectiveness of our proposed algorithm is demonstrated through computational experiments. Of the tested 30 real-world instances provided by a company, 28 obtained outcomes are as good as the arrangements designed by human experts.

ABSTRACT. This study deals with cutting tool management in flexible manufacturing systems (FMS). Typically, FMS comprise multi-axis machine tools equipped with automated tool changers, material-handling robots, and large-capacity tool magazines. We propose an optimization model that addresses the increased use of various cutting tools and improves parallel-machine load balancing. This study builds on our previous work by introducing an optimization model for allocating cutting tools to CNC machine tools within FMS, considering balanced workloads among machine tools and the costs associated with cutting tool utilization. The study investigates how tool allocation and cutting tool costs influence production rates through computational experiments.

ABSTRACT. Many social infrastructures in Japan are deteriorating. In this study, autonomous hammering robot was developed to replace the infrastructure inspections performed by inspectors. The autonomous hammering robot con-sists of a mobile unit that moves, an arm unit that determines the position of its hand, and an inspection unit that inspects the wall surface by hammering and collecting sound. Scheduling tunnel inspections and sound inspections of the entire infrastructure were confirmed to be possible with the developed robot.

ABSTRACT. Recent advances in AI and IoT have spurred smart agriculture innovations. This study developed a smart water-supply device that schedules oral temperature measurements in pigs while drinking. The system uses a switch activated when pigs bite the water and temperature measurement pipes, ensuring measurements occur before water inflow. Evaluations showed that delay improves measurement accuracy by reducing water temperature influence, and the highest oral temperatures moderately correlate (0.55) with vaginal temperatures, suggesting potential as a proxy for core body temperature.

ABSTRACT. With the rapid increase in IP traffic due to the advancement of the global information society, elastic optical networks, which flexibly allocate bandwidth for connection requests, are becoming increasingly important. Unlike traditional optical networks, elastic optical networks improve resource utilization efficiency. However, popular path selection methods, which prioritize shorter physical paths or fewer hops, face challenges such as handling concentrated connection requests and ensuring path diversity. This study proposes new adaptive path selection methods based on communication demand and network characteristics. Experiments show that proposed methods allocates optical resources more efficiently than traditional methods.

ABSTRACT. Given the locations of n airports and the revenues between each pair of airports in advance, the point-to-point airline network design problem seeks an optimal location of m edges to maximize a company's revenue. This paper presents a constructive and three local search heuristics to obtain high-quality solutions for large-scale instances (up to n=100). This paper reports the performance of the heuristics from the perspectives of the computation time and the approximation ratio. We observe that the constructive algorithm produces high-quality solutions, and local search heuristics often improve the solutions further.

ABSTRACT. There is a worldwide movement toward pedestrian-friendly urban development, driven by expectations of lower environmental impact, reduced roadway space usage, and health benefits associated with physical activity. In Japan, however, it is also essential to develop safe and comfortable pedestrian spaces that remain accessible even during increasingly severe summer heat. We present two network design problems for optimizing the placement of facilities such as shade-providing street trees and arcades to enhance pedestrian safety. The first focuses on minimizing total social cost, while the second aims to maximize the number of pedestrians who can travel safely within heat exposure limits.

ABSTRACT. This paper proposes an optimized algorithm for bicycle painting scheduling, addressing the bottleneck status of the painting station due to the diversity of products and paints, as well as production constraints. To overcome challenges such as production delays, frequent color changes, and excessive handling caused by traditional manual scheduling, we developed an innovative scheduling method based on a genetic algorithm. This method automatically optimizes the production process, significantly reducing completion time, minimizing color change occurrences, and decreasing handling frequency, thereby substantially enhancing production efficiency. Experimental results validate the method’s effectiveness, demonstrating its applicability across multiple scenarios. Future research will focus on expanding the application scope and continuously optimizing the performance of the simulation program.

ABSTRACT. The purpose of this paper is to develop an optimal production scheduling model for the vulcanization process and an algorithm to derive the optimal solution. Furthermore we attempt to conduct an empirical case analysis using data from a manufacturing company in Korea. In particular, this study proposes a practical production scheduling approach that reflects the relationships among products, molds, and vulcanizers within the vulcanization process. Efforts were made to present an algorithm capable of generating solutions in real time to ensure applicability in real-world settings.

ABSTRACT. This study proposes employing a Decision Transformer (DT) to tackle the Job-Shop Scheduling Problem (JSSP) via offline reinforcement learning on a diverse dataset. The DT reformulates scheduling tasks as sequential decision-making processes and leverages a GPT-style Transformer architecture to process trajectories that comprise Return-to-Go, states, and actions. We carefully design the state, action, and reward components and employ various strategies to collect training trajectories for model training. Experimental results indicate that the DT significantly outperforms classical Priority Dispatching Rules in terms of makespan, suggesting superior adaptability and robustness in scheduling tasks.