Modelling and symmetry breaking in scheduling problems on batch processing machines

Problems of scheduling batch-processing machines to minimise the makespan are widely exploited in the literature, mainly motivated by real-world applications, such as burn-in tests in the semiconductor industry. These problems consist of grouping jobs in batches and scheduling them on machines. We consider problems where jobs have non-identical sizes and processing times, and the total size of each batch cannot exceed the machine capacity. The processing time of a batch is defined as the longest processing time among all jobs assigned to it. Jobs can also have non-identical release times, and in this case, a batch can only be processed when all jobs assigned to it are available. This paper discusses four different versions of batch scheduling problems, considering a single processing machine or parallel processing machines and considering jobs with or without release times. New mixed integer linear programming formulations are proposed as enhancements of formulations proposed in the literature, and symmetry breaking constraints are investigated to reduce the size of the feasible sets. Computational results show that the proposed formulations have a better performance than other models in the literature, being able to solve to optimality instances only considered before to be solved by heuristic procedures.     

Probabilistic asymptotic analysis of stochastic online scheduling problems

In the stochastic online scheduling environment, jobs with unknown release times and weights arrive over time. Upon arrival, the information on the weight of the job is revealed but the processing requirement remains unknown until the job is finished. In this paper we consider the objective of minimizing the total weighted completion time. With the assumptions that job weights are bounded, machine capacity is adequate, and processing requirements are bounded and identical and independently distributed across the machines and jobs, we show that any nondelay algorithm is asymptotically optimal for the stochastic online single machine problem, flow shop problem, and uniform parallel machine problem. Our simulation studies of these stochastic online scheduling problems show that two generic nondelay algorithms perform very well as long as the number of jobs is larger than 100.     

Minimising makespan on parallel batch processing machines with non-identical ready time and arbitrary job sizes

This paper considers the parallel batch processing machine scheduling problem which involves the constraints of unequal ready times, non-identical job sizes, and batch dependent processing times in order to sequence batches on identical parallel batch processing machines with capacity restrictions. This scheduling problem is a practical generalisation of the classical parallel batch processing machine scheduling problem, which has many real-world applications, particularly, in the aging test operation of the module assembly stage in the manufacture of thin film transistor liquid crystal displays (TFT-LCD). The objective of this paper is to seek a schedule with a minimum total completion time for the parallel batch processing machine scheduling problem. A mixed integer linear programming (MILP) model is proposed to optimise the scheduling problem. In addition, to solve the MILP model more efficiently, an effective compound algorithm is proposed to determine the number of batches and to apply this number as one parameter in the MILP model in order to reduce the complexity of the problem. Finally, three efficient heuristic algorithms for solving the large-scale parallel batch processing machine scheduling problem are also provided.     

含不一致任务重量的同型熔炼炉批调度优化

为求解含不一致任务重量的同型熔炼炉批调度问题,建立了最小化最大任务完工时间优化模型,设计了一种混合粒子群算法（HPSO）。算法使用随机生成的任务序列作为粒子,采用批首次匹配（BFF）规则对任务序列分批,最长加工时间（LPT）规则将批分配到批处理机,并提出了一种最小完工时间差（MCD）规则对LPT调度结果进行优化;为避免早熟,算法引入交叉和变异操作搜索最优解。通过仿真实验与SA、GA算法对比,实验结果表明算法具有良好的性能。     

Minimizing the makespan for the two-machine scheduling problem with a single server: Two algorithms for very large instances

This article considers the problem of scheduling a given set of n jobs on two identical parallel machines with a single server. Each job must be processed on one of the machines. Before processing, the server has to set up the relevant machine. The objective is to minimize the makespan. For this unary NP-hard problem, two fast constructive algorithms with a complexity of O(n²) are presented. The performance of these algorithms is evaluated for instances with up to 10,000 jobs. Computational results indicate that the algorithms have an excellent performance for very large instances so that the obtained objective function values are very close to a lower bound, and in many cases even an optimal solution is achieved. Superiority over all existing algorithms is obtained by sequencing the jobs on the two machines so that the machine idle time and the server waiting time are minimized. In doing so, the characteristics of an optimal solution resulting from its relevant lower bound are taken into account.     

Minimizing makespan in a two-stage flow shop with parallel batch-processing machines and re-entrant jobs

Against a background of heat-treatment operations in mould manufacturing, a two-stage flow-shop scheduling problem is described for minimizing makespan with parallel batch-processing machines and re-entrant jobs. The weights and release dates of jobs are non-identical, but job processing times are equal. A mixed-integer linear programming model is developed and tested with small-scale scenarios. Given that the problem is NP hard, three heuristic construction methods with polynomial complexity are proposed. The worst case of the new constructive heuristic is analysed in detail. A method for computing lower bounds is proposed to test heuristic performance. Heuristic efficiency is tested with sets of scenarios. Compared with the two improved heuristics, the performance of the new constructive heuristic is superior.     

JOB ASSIGNMENT ALGORITHMS ON UNIFORM PARALLEL MACHINES

Uniform parallel machine scheduling problems with a makespan measure cannot generally be solved within polynomial time complexity. This paper considers special problems with a single type of job on the uniform parallel machines, where each machine is available at a given ready time. Also the machine can be restricted on the number of jobs to be processed. The objective is to develop job assignment or batching algorithms which minimize makespan. When all the machines are available at time zero and have no restriction on the number of assignable jobs, a lower bound and optimal solution properties are derived. Based upon these properties, a polynomial algorithm is suggested to find the optimal job assignment on each machine. Three generalized problems are considered under the following situations: (1) some machines have capacity restrictions on the production batch, (2) each machine has its ready time, and (3) the jobs require series-parallel operations. The generalized problems arc also characterized and polynomial algorithms are developed for the same aim of optimal job assignment, except for the case of series-parallel operations. A heuristic algorithm is suggested with numerical tests for the series-parallel operations problem     

Minimising interference for scheduling two parallel machines with a single server

In this paper, we consider the problem of scheduling a set of jobs on two parallel machines with set-up times. The set-up has to be performed by a single server. The objective is to minimise the forced idle time. The problem of minimising the forced idle time (interference problem) is known to be unary NP-hard for the case of two machines and equal set-up and arbitrary processing times. We propose a mixed integer linear programming model, which describes a special class of schedules where the jobs from a list are scheduled alternatively on the machines, and a heuristic algorithm is tested on instances with up to 100,000 jobs. The computational results indicate that the algorithm has an excellent performance even for very large instances, where mostly an optimal solution is obtained within a very small computational time.     

The Value of Processing Flexibility in Multipurpose Machines

We study a scheduling problem in a multipurpose machine environment where every job can be processed by a subset of the machines operated in parallel, with the objective of minimizing makespan. We develop lower bounds, heuristic algorithms, and a branch-and-bound procedure. We perform an extensive computational experiment to assess how much flexibility is enough to render the multipurpose machine system equally efficient to an equivalent system of parallel identical machines. We find that very small amounts of flexibility appropriately distributed across processors provide nearly the same makespan performance as a system of fully flexible parallel machines.     

Two-stage hybrid flow shop scheduling with preventive maintenance using multi-objective tabu search method

This paper investigates an integrated bi-objective optimisation problem with non-resumable jobs for production scheduling and preventive maintenance in a two-stage hybrid flow shop with one machine on the first stage and m identical parallel machines on the second stage. Sequence-dependent set-up times and preventive maintenance (PM) on the first stage machine are considered. The scheduling objectives are to minimise the unavailability of the first stage machine and to minimise the makespan simultaneously. To solve this integrated problem, three decisions have to be made: determine the processing sequence of jobs on the first stage machine, determine whether or not to perform PM activity just after each job, and specify the processing machine of each job on the second stage. Due to the complexity of the problem, a multi-objective tabu search (MOTS) method is adapted with the implementation details. The method generates non-dominated solutions with several parallel tabu lists and Pareto dominance concept. The performance of the method is compared with that of a well-known multi-objective genetic algorithm, in terms of standard multi-objective metrics. Computational results show that the proposed MOTS yields a better approximation.     

A computational study of heuristics for two-stage flexible flowshops

We consider the scheduling of two-stage flexible flowshops. This manufacturing environment involves two machine centres representing two consecutive stages of production. Each machine centre is composed of multiple parallel machines. Each job has to be processed serially through the two machine centres. In each machine centre, a job may be processed on any of the machines. There are n independent jobs to be scheduled without preemption. The jobs can wait in between the two machine centres and the intermediate storage is unlimited. Our objective will be to minimize the maximum completion time of the jobs. We formulate the problem as a mixed integer program. Given this problem class is NP-hard in the strong sense, we present three lower bounds to estimate the optimal solution. We then propose a sequence-first, allocate-second heuristic approach for its solution. We heuristically decompose the problem by first creating a priority list to order the jobs and then assign the jobs to the available machines in each machine centre based on this order. We describe seven rules for the sequencing phase. The assignment phase consists of a heuristic which attempts to minimize each partial schedule length while looking ahead at the future assignment of the currently unscheduled jobs. The computational performance of the heuristic approach was evaluated by comparing the value of each heuristic variant to the best among the three lower bounds. Its effectiveness was tested on scenarios pertinent to flexible flowshop environments, such as cellular manufacturing, by conducting a computational study of over 3400 problems. Our computational results indicate that the most effective approach used Johnson's rule to provide the priority list for job assignment. This provided integrality gaps which on the average were less than 0·73%.     

Spread time considerations in operational fixed job scheduling

In this study, we consider the operational fixed job scheduling problem on identical parallel machines. We assume that the jobs have fixed ready times and deadlines, and spread time constraints are imposed on machines. Our objective is to select a set of jobs for processing so as to maximise the total weight. We show that the problem is strongly NP-hard, and we investigate several special polynomially solvable cases. We propose a branch and bound algorithm that employs size reduction mechanisms, dominance conditions, and powerful lower and upper bounds. The computational results reveal that the branch and bound algorithm returns optimal solutions for problem instances with up to 100 jobs in reasonable solution times.     

Towards robustness of response times: minimising the maximum inter-completion time on parallel machines

Motivated by scheduling practices that require a response to unplanned high-priority jobs as soon as possible without preempting any in-processing jobs, this paper considers a deterministic identical parallel machine scheduling problem to achieve robustness with regard to a worst-case response time. To the best of our knowledge, this paper is the first to study the objective of minimising the maximum inter-completion time, i.e. the maximum time difference between any two consecutive completion times of jobs. For this novel scheduling problem, we first show its NP-hardness, and then propose an integer linear programming formulation and three heuristic approaches. Numerical experiments demonstrate the efficiency of our solution methods.     

Knapsack-Based Approaches to the Makespan Problem on Multiple Processors

In this paper we study the problem of scheduling n independent jobs available at time zero on m ≥ 2 parallel and identical processors with the objective of minimizing the makespanJWe propose two approaches, both are Knapsack-based. The first is analytical and depends on reducing the set of m occupancy constraints to a single Diophantine equation. We capitalize on the very special structure of the ILP to effect the reduction with small multipliers. The second approach is an iterative heuristic procedure that is based on the observation that a two-machine makespan problem is trivially reduced to a Knapsack problem. Computational experience indicates the superiority of this approach over other existing approaches. Realistic problems of up to 100 jobs on 10 machines are solved in a few seconds on the IBM 370/165.     

Minimizing makespan on parallel machines with batch arrivals

Most studies in the scheduling literature assume that jobs arrive at time zero, while some studies assume that jobs arrive individually at non-zero times. However, both assumptions may not be valid in practice because jobs usually arrive in batches. In this article, a scheduling model for an identical parallel machine problem with batch arrivals is formulated. Because of the NP-hardness of the problem, a heuristic based on a simplified version of lexicographical search is proposed. To verify the heuristic, two lower bounding schemes are developed, where one lower bound is tight, and the list scheduling heuristic is compared. Extensive computational experiments demonstrate that the proposed heuristic is quite efficient in obtaining near optimal solution with an average error of less than 1.58%. The percentage improvement (from the lower bound) of the heuristic solution on the solution by the list scheduling is as large as 31.68.     

Unrelated parallel machine scheduling with job splitting

Scheduling jobs on unrelated parallel machines is an activity that is very much a part of industrial scheduling. We report a methodology for minimizing the total weighted tardiness of all jobs intended to be processed on unrelated parallel machines in the presence of dynamic job releases and dynamic machine availability. More importantly, the mixed (binary) integer linear programming model formulated for the problem incorporates a couple of “hard” operational constraints to ensure that just-in-time manufacturing practices are followed by controlling the work-in-process and/or finished goods inventories generated by split jobs mandated by a tight due date, a high priority, and/or a high workload. Four different methods based on simple and composite dispatching rules are used to identify an initial solution, which is then used by the tabu-search-based heuristic solution algorithm to ultimately find the best solution. Incorporating the various tabu search features led to the development of six different heuristics that were tested on eight small problem instances to compare the quality of their solutions to the optimal solutions. The results show that the proposed heuristics are capable of obtaining solutions of good quality in a remarkably short computation time with the best performer among them recording a percentage deviation of only 1.18%. A factorial experiment based on a split-plot design is performed to test the performance of the heuristics on problem structures, ranging from nine jobs and three machines to 60 jobs and 15 machines. The results show that the newly developed composite dispatching heuristic, referred to as the modified apparent tardiness cost, is capable of obtaining initial solutions that significantly accelerate the tabu-search-based heuristics to attain the best solution. The use of a long-term memory function is proven to be advantageous in solving all problem structures. In addition, the variable tabu list size is preferred for solving the small problem structure, while the fixed tabu list size is preferred as the problem size grows from small to medium and then large.     

Sequencing n Products Involving m Independent Jobs on m Machines

This article considers the problem of scheduling n products over m distinct machines. Every product consists of a set of jobs, each requiring a known processing time on a designated machine. There are no precedence constraints, and simultaneous processing of jobs requiring different machines within a product is allowed. The object of scheduling is to minimize a regular measure of performance associated with the products. It is shown that there exists an optimal schedule with the “no passing property.” Branch and bound routines are developed for finding the optimal solution for the two measures of performance: (1) total penalty cost; and (2) sum of product completion times. Comparisons between the optimal solution and solutions obtained using dispatching rules are given in the penalty cost case.     

Scheduling parallel batch processing machines with arbitrary job sizes and incompatible job families

Motivated by a bottleneck operation in an MLCC (multi-layer ceramic capacitor) production line, we study the scheduling problem of parallel batch processing machines in which a number of jobs can be processed simultaneously in a machine as a batch. Volumes of the jobs are different from each other and each job belongs to the family in which all jobs have the same processing time. In this situation, we analyse three kinds of problems whose performance measures are makespan, total completion time, and total weighted completion time, respectively. Since these problems are known to be NP-hard, we propose a number of heuristics and design genetic algorithms for the problems. Through some computational experiments, we evaluate the performances of the heuristic algorithms proposed, including the genetic algorithms for each of three problems.     

A fixed job scheduling problem with machine-dependent job weights

This study considers the identical parallel machines operational fixed job scheduling problem with machine-dependent job weights. A job is either processed in a fixed interval or is not processed at all. Our aim is to maximise the total weight of the processed jobs. We show that the problem with machine eligibility constraints resides as a special case of this problem. We identify some special polynomially solvable cases and propose a branch-and-bound (BB) algorithm that employs efficient bounding schemes and dominance conditions. Computational experience on large-sized problem examples reveals the satisfactory performance of the BB algorithm.     

On-line scheduling of multi-server batch operations

The batching of jobs in a manufacturing system is a very common policy in many industries. The main reasons for batching are the avoidance of setups and/or facilitation of material handling. Good examples of batch-wise production systems are the ovens that are found in the aircraft industry and in the manufacture of semiconductors. These systems often consist of multiple machines of different types for the range and volumes of products that have to be handled. Building on earlier research in the aircraft industry, where the process of hardening synthetic aircraft parts was studied, we propose a new heuristic for the dynamic scheduling of these types of systems. Our so-called look-ahead strategy bases its decision to schedule a job on a certain machine on the availability of information on a limited number of near future arrivals. The new control strategy distinguishes itself from existing heuristics by an integrated approach that involves all machines in the scheduling decision, instead of only considering idle machines. It is shown by an extensive series of simulation experiments that the new heuristic outperforms existing heuristics for most system configurations. Especially in the case of complex systems, where multiple products have to be handled by non-identical machines, the new heuristic proves its value as a practical scheduling tool. Important insight is obtained with regard to the relation between the system is configuration and its performance.