Liji Shen

Solving the serial batching problem in job shop manufacturing systems

This paper addresses the serial batch scheduling problem embedded in a job shop environment to minimize makespan. Sequence dependent family setup times and a job availability assumption are also taken into account. In consideration of batching decisions, we propose a tabu search algorithm which consists of various neighborhood functions, multiple tabu lists and a sophisticated diversification structure. Computational experiments show that our algorithm outperforms a well-known tabu search approach which is developed for solving the traditional job shop problem. These results also confirm the benefits of batching.

An integrated tabu search algorithm for the lot streaming problem in job shops

In this paper, we focus on solving the lot streaming problem in a job shop environment, where consistent sublots are considered. The presented three-phase algorithm incorporates the predetermination of sublot sizes, the determination of schedules based on tabu search and the variation of sublot sizes. With regard to tabu search implementation, a constructive multi-level neighbourhood is developed, which effectively connects three isolated neighbourhood functions. Moreover, enhancements of the basic version of tabu search are conducted. Combined with the procedure for varying sublot sizes, the algorithm further exploits the improvement potential. All tested instances show a rapid convergence to their lower bounds. The well-known difficult benchmark problems also achieve substantial makespan reduction. In addition, the performance of specific components is intensively examined in our study.

A study of hybrid evolutionary algorithms for single machine scheduling problem with sequence-dependent setup times

We present a systematic comparison of hybrid evolutionary algorithms (HEAs), which independently use six combinations of three crossover operators and two population updating strategies, for solving the single machine scheduling problem with sequence-dependent setup times. Experiments show the competitive performance of the combination of the linear order crossover operator and the similarity-and-quality based population updating strategy. Applying the selected HEA to solve 120 public benchmark instances of the single machine scheduling problem with sequence-dependent setup times to minimize the total weighted tardiness widely used in the literature, we achieve highly competitive results compared with the exact algorithm and other state-of-the-art metaheuristic algorithms in the literature. Meanwhile, we apply the selected HEA in its original form to deal with the unweighted 64 public benchmark instances. Our HEA is able to improve the previous best known results for one instance and match the optimal or the best known results for the remaining 63 instances in a reasonable time.

Flow shop batching and scheduling with sequence-dependent setup times

This paper addresses the flow shop batching and scheduling problem where sequence-dependent family setup times are present and the objective is to minimize makespan. We consider violating the group technology assumption by dividing product families into batches. In order to reduce setup times, inconsistent batches are formed on different machines, which lead to non-permutation schedules. To the best of our knowledge, this is the first time that the splitting of job families into inconsistent batches has been considered in a flow shop system. A tabu search algorithm is developed which contains several neighbourhood functions, double tabu lists and a multilevel diversification structure. Compared to the state-of-the-art meta-heuristics for this problem, the proposed tabu search algorithm achieves further improvement when the group scheduling assumption is dropped. Also, various experiments conducted on the benchmark problem instances confirm the benefits of batching. Therefore, it will be prudent for the practitioners to consider adopting inconsistent batches and non-permutation schedules to improve their operational efficiency within a reasonable amount of computational effort.

An iterative approach for the serial batching problem with parallel machines and job families

In this paper, we consider a parallel machine scheduling problem to minimize the total completion time. Each job belongs to a certain family. All jobs of one family have identical processing times. Major setups occur between jobs of different families, and we include sequence dependencies. Batches of jobs belonging to the same family can be formed to avoid these setups. Furthermore, we assume serial batching and batch availability. Therefore, the processing time of a batch is the sum of the processing times of all jobs grouped into the corresponding batch. An iterative method is developed for solving this specific problem. This approach alternates between varying batch sizes using an efficient heuristic and sequencing batches based on variable neighborhood search (VNS). Computational results demonstrate that the iterative heuristic outperforms heuristics based on a fixed batch size and list scheduling.

A simultaneous and iterative approach for parallel machine scheduling with sequence-dependent family setups

In this paper, we address a parallel machine scheduling problem to minimize the total weighted completion time, where product families are involved. Major setups occur when processing jobs of different families, and sequence dependencies are also taken into account. Considering its high practical relevance, we focus on the special case where all jobs of the same family have identical processing times. In order to avoid redundant setups, batching jobs of the same family can be performed. We first develop a variable neighborhood search algorithm (VNS) to solve the interrelated subproblems in a simultaneous manner. To further reduce computing time, we also propose an iterative scheme which alternates between a specific heuristic to form batches and a VNS scheme to schedule entire batches. Computational experiments are conducted which confirm the benefits of batching. Test results also show that both simultaneous and iterative approach outperform heuristics based on a fixed batch size and list scheduling. Furthermore, the iterative procedure succeeds in balancing solution quality and computing time.

Solving the flexible job shop scheduling problem with sequence-dependent setup times

This paper addresses the flexible job shop scheduling problem with sequence-dependent setup times and where the objective is to minimize the makespan. We first present a mathematical model which can solve small instances to optimality, and also serves as a problem representation. After studying structural properties of the problem using a disjunctive graph model, we develop a tabu search algorithm with specific neighborhood functions and a diversification structure. Extensive experiments are conducted on benchmark instances. Test results first show that our algorithm outperforms most existing approaches for the classical flexible job shop scheduling problem. Additional experiments also confirm the significant improvement achieved by integrating the propositions introduced in this study.

MIP formulations and heuristics for solving parallel batching problems

This paper addresses the scheduling problem involving batch processing machines, which is also known as parallel batching in the literature. The presented mixed integer programming formulation first provides an elegant model for the problem under study. Furthermore, it enables solutions to the problem instances beyond the capability of exact methods developed so far. In order to alleviate computational burden, the authors propose MIP-based heuristic approaches which balance solution quality and computing time.

Family scheduling with batch availability in flow shops to minimize makespan

This paper addresses a batch scheduling problem in flow shop production systems, where job families are formed based on setup similarities. In order to improve setup efficiency, we consider batching decisions in our solution procedure. Due to its high practical relevance, the batch availability assumption is also adopted in this study. In the presence of sequence-dependent setup times, it is proved that a permutation flow shop is generally not optimal. Therefore, our objective is to determine solutions with inconsistent batches, which essentially lead to non-permutation schedules, to minimize makespan. After examining structural properties, we develop a tabu search algorithm with multiple neighbourhood functions. Computational results confirm the remarkable benefits of batching decisions. Our algorithm also outperforms some well-known and well-performing approaches.

Solving the Batch Scheduling Problem with Family Setup Times

In this paper we address the machine scheduling problem involving family setup times and batching decisions. The m-machine ow shop system is considered with the objective of minimizing the makespan. We first present a mathematical formulation which is able to solve small instances. Subsequently, a tabu search algorithm is proposed with diverse neighbourhoods.