Explore the vast range of titles available.

Order scheduling is of a great significance in the internet and communication industries. With the rapid development of the communication industry and the increasing variety of user demands, the number of work orders for communication operators has grown exponentially. Most of the research that tries to solve the order scheduling problem has focused on improving assignment rules based on real-time performance. However, these traditional methods face challenges such as poor real-time performance, high human resource consumption, and low efficiency. Therefore, it is crucial to solve multi-objective problems in order to obtain a robust order scheduling policy to meet the multiple requirements of order scheduling in real problems. The priority dispatching rule (PDR) is a heuristic method that is widely used in real-world scheduling systems In this paper, we propose an approach to automatically optimize the Priority Dispatching Rule (PDR) using a deep multiple-objective reinforcement learning agent and to optimize the weighted vector with a convex hull to obtain the most objective and efficient weights. The convex hull method is employed to calculate the maximal linearly scalarized value, enabling us to determine the optimal weight vector objectively and achieve a balanced optimization of each objective rather than relying on subjective weight settings based on personal experience. Experimental results on multiple datasets demonstrate that our proposed algorithm achieves competitive performance compared to existing state-of-the-art order scheduling algorithms.

The application of nature-inspired algorithms and priority hybrid dispatching rules for simultaneous scheduling and dispatching of automatic guided vehicles (AGVs) are observed to be highly significant for the best utilization of the flexible manufacturing system (FMS) facility. The earlier studies on the use of AGVs in the FMS have generally focused on minimizing the complexity of AGV operations by optimizing their material handling schedule and their routing in different types of FMS configurations. However, this is achieved only by using an appropriate optimizing algorithm and dispatching rule under different FMS operating conditions. The aim of this paper is to use the simulation methodology so as to compare and analyze the combined effect of four experimental factors, namely four types of priority hybrid dispatching rules, three different nature-inspired algorithms, two levels of loading/unloading times and two levels of machine failures, on the different performance parameters of the FMS. The performance parameters of the FMS analyzed are simultaneous minimization in distance travel and backtracking of AGV, the total production rate of the FMS, mean AGV utilization and mean work center utilization in the FMS. Additionally, in order to find the mean and interaction effect of the experimental factors on the aforementioned performance parameters of the FMS, an analysis of variance (ANOVA) is also carried out. From the results, it is observed that the interaction between experimental factors, namely nature-inspired algorithms and loading–unloading time, has a significant effect on the performance measures, namely simultaneous reduction in distance travel and backtracking, mean work center utilization (%) and total production rate of the FMS. The interaction between aforesaid experimental factors has no significant effect on mean AGV utilization (%) in the FMS. However, the interaction between priority hybrid dispatching rules and loading/unloading times is found to have significant effect on the mean work center utilization (%) in the FMS facility.

The flexible manufacturing system (FMS) constitute of several programmable production work centers, material handling systems (MHSs), assembly stations and automatic storage and retrieval systems. In FMS, the automatic guided vehicles (AGVs) play a vital role in material handling operations and enhance the performance of the FMS in its overall operations. To achieve low makespan and high throughput yield in the FMS operations, it is highly imperative to integrate the production work centers schedules with the AGVs schedules. The Production schedule for work centers is generated by application of the Giffler and Thompson algorithm under four kind of priority hybrid dispatching rules. Then the clonal selection algorithm (CSA) is applied for the simultaneous scheduling to reduce backtracking as well as distance travel of AGVs within the FMS facility. The proposed procedure is computationally tested on the benchmark FMS configuration from the literature and findings from the investigations clearly indicates that the CSA yields best results in comparison of other applied methods from the literature.

Selection of appropriate priority dispatching rules (PDRs) is a major concern in practical scheduling problems. Earlier research implies that using one PDR does not necessarily yield to an optimal schedule. Hence, this paper puts forward a novel approach based on discrete event simulation (DES) and artificial neural networks (ANNs) to decide on the optimal PDR for each machine from a set of rules so as to minimize the makespan in job shop scheduling problems. Non-identical PDRs are considered for each machine. Indeed, for a given number of machines, all permutations of PDRs are taken into account which could lead to nondeterministic polynomial-time hardness of the problem when the number of machines increases. To address this issue, DES and ANNs are employed as a meta-model. First, the problem is modeled and quite a number of feasible solutions are obtained from DES on its own. Afterward, a back-propagation neural network is developed in accordance with the results of DES to calculate the makespan based on all potential permutations of PDRs. The performance of the proposed approach is investigated through a set of test-bed problems.

ABSTRACT Cloud‐edge collaborative manufacturing supports cloud‐based training and edge‐side execution, requiring production scheduling policies with both rapid response and strong generalization. For the flexible job shop scheduling problem, existing deep reinforcement learning (DRL) methods usually rely on priority dispatching rules which limit generalization performance. To address this issue, this paper proposes heterogeneous graph attention‐enhanced DRL‐based scheduling (HGA‐DS), a heterogeneous graph attention‐enhanced deep reinforcement learning framework, and adopts dynamic joint operation‐machine selection for scheduling decisions. Specifically, an extended heterogeneous graph with job nodes is introduced to improve state representation, an estimated maximum finish time based reward function is designed to provide more effective training feedback, and a lightweight state‐conditioned attention mechanism is developed for low‐overhead neighbourhood information aggregation. Experimental results on datasets of different scales show that HGA‐DS outperforms existing learning‐based methods and composite dispatching rules, while exhibiting strong generalization ability and real‐time decision‐making performance.

Assembly job shops form an important part of make-to-order companies. These high-variety production environments are generally characterized by high shop loads and tight delivery dates. Coordinating the completion times of parts to guarantee a timely start of the assembly operations is complex. Most prior studies on coordination in assembly job shops mainly focus on priority dispatching rules, thereby neglecting the coordinating potential of order release decisions. While release and dispatching methods have been extensively studied in the literature, they lack the refined dynamic mechanisms that are needed to effectively coordinate assembly parts. This study refines order pool sequencing rules for order release by utilizing progress, urgency and load-related status information for coordination purposes. A new selection mechanism focusing on the timely release of critical parts is embedded in the release decision. Furthermore, a newly developed dynamic dispatching rule carefully coordinates parts to be assembled, once they are released to the shop floor. Simulation results show that the newly developed methods for dynamic coordination significantly outperform their static versions.

As the focus on environmental sustainability sharpens, the significance of low-carbon manufacturing and energy conservation continues to rise. While traditional flexible job shop scheduling strategies are primarily concerned with minimizing completion times, they often overlook the energy consumption of machines. To address this gap, this paper introduces a novel solution utilizing deep reinforcement learning. The study begins by defining the Low-carbon Flexible Job Shop Scheduling problem (LC-FJSP) and constructing a disjunctive graph model. A sophisticated representation, based on the Markov Decision Process (MDP), incorporates a low-carbon graph attention network featuring multi-head attention modules and graph pooling techniques, aimed at boosting the model’s generalization capabilities. Additionally, Bayesian optimization is employed to enhance the solution refinement process, and the method is benchmarked against conventional models. The empirical results indicate that our algorithm markedly enhances scheduling efficiency by 5% to 12% and reduces carbon emissions by 3% to 8%. This work not only contributes new insights and methods to the realm of low-carbon manufacturing and green production but also underscores its considerable theoretical and practical implications.

Selection hyper-heuristics are novel tools that combine low-level heuristics into robust solvers commonly used for tackling combinatorial optimization problems. However, the training cost is a drawback that hinders their applicability. In this work, we analyze the effect of training with different problem sizes to determine whether an effective simplification can be made. We select Job Shop Scheduling problems as an illustrative scenario to analyze and propose two hyper-heuristic approaches, based on Simulated Annealing (SA) and Unified Particle Swarm Optimization (UPSO), which use a defined set of simple priority dispatching rules as heuristics. Preliminary results suggest a relationship between instance size and hyper-heuristic performance. We conduct experiments training on two different instance sizes to understand such a relationship better. Our data show that hyper-heuristics trained in small-sized instances perform similarly to those trained in larger ones. However, the extent of such an effect changes depending on the approach followed. This effect was more substantial for the model powered by SA, and the resulting behavior for small and large-sized instances was very similar. Conversely, for the model powered by UPSO, data were more outspread. Even so, the phenomenon was noticeable as the median performance was similar between small and large-sized instances. In fact, through UPSO, we achieved hyper-heuristics that performed better on the training set. However, using small-sized instances seems to overspecialize, which results in spread-out testing performance. Hyper-heuristics resulting from training with small-sized instances can outperform a synthetic Oracle on large-sized testing instances in about 50% of the runs for SA and 25% for UPSO. This allows for significant time savings during the training procedure, thus representing a worthy approach.

In today’s complex and unpredictable manufacturing environment, dynamic events such as new job arrivals and urgent insertions can occur at any time. Meanwhile, multiple conflicting objectives need to be optimized simultaneously in the flexible job-shop scheduling problem (FJSP). This necessitates real-time multi-objective FJSP scheduling methods that can balance time efficiency and solution quality. Therefore, this paper proposes a dynamic multi-objective FJSP method based on a dynamic dual-attention network (DDAN) and multi-agent reinforcement learning. The DDAN captures global feature representations of operations and machines and deeply explores their complex dependencies. By integrating dynamic attention coefficients with job urgency factors, the DDAN can respond in real-time to dynamic events and provide effective support for subsequent decision-making. Additionally, a multi-agent reinforcement learning framework is introduced to balance the conflicting objectives of makespan and average tardiness. The higher level agent is designed to optimize makespan, while the lower level agent focuses on minimizing average tardiness. Collaboration between the two agents is facilitated by a carefully designed state-sharing mechanism and distinct reward functions, enabling more flexible and efficient resolution of conflicts in multi-objective optimization. Extensive testing has demonstrated the exceptional performance of the proposed method, which consistently and rapidly converges to the optimal solution, outperforming traditional priority dispatching rules (PDRs), metaheuristic algorithms, and state-of-the-art reinforcement learning methods, particularly in handling large-scale test instances.

This paper shows the effectiveness of labour transfers in addressing premature idleness caused by controlled order release. Controlled order release restricts order entry to the shop floor and is commonly employed in high-variety manufacturing where it results in benefits such as stable work-in-progress. However, it can increase waiting times when orders are blocked from release, while capacities are idling. This issue, known as premature idleness, negatively impacts delivery performance. Previous studies have primarily focused on addressing premature idleness through input control by releasing new orders to idling workstations. This approach overlooks the potential of output control during premature idleness, transferring labour to assist at other workstations in a dual resource constrained setting. Using simulation, this study demonstrates that output control significantly improves delivery performance—in terms of mean tardiness and percentage tardy—and reduces total and shop floor throughput times. Importantly, this result proves robust, even when the efficiency of the assisting worker is severely limited. Shop-level performance improves despite the efficiency loss of the worker. The impact of the where-rule is minimal, while the efficacy of the priority dispatching rule depends on the joint efficiency of collaborating workers. Finally, we show that combining input control and output control enhances performance, providing opportunities for further research on the role of both control approaches in high-variety manufacturing.