Interactive optimization

Optimization algorithms or heuristics in which the user interacts significantly either during the solution process or as part of post-optimality analysis are becoming increasingly popular. An important underlying premise of such man/machine systems is that there are some steps in solving a problem in which the computer has an advantage and other steps in which a human has an advantage. This paper first discusses how man/machine systems can be useful in facilitating model specification and revision, coping with aspects of a problem that are difficult to quantify and assisting in the solution process. We then survey successful systems that have been developed in the areas of vehicle scheduling, location problems, job shop scheduling, course scheduling, and planning language-based optimization.     

Exploiting process plan flexibility in production scheduling: A multi-objective approach

When solving a product/process design problem, we must exploit the available degrees of freedom to cope with a variety of issues. Alternative process plans can be generated for a given product, and choosing one of them has implications on manufacturing functions downstream, including planning/scheduling. Flexible process plans can be exploited in real time to react to machine failures, but they are also relevant for off-line scheduling. On the one hand, we should select a process plan in order to avoid creating bottleneck machines, which would deteriorate the schedule quality; on the other one we should aim at minimizing costs. Assessing the tradeoff between these possibly conflicting objectives is difficult; actually, it is a multi-objective problem, for which available scheduling packages offer little support. Since coping with a multi-objective scheduling problem with flexible process plans by an exact optimization algorithm is out of the question, we propose a hierarchical approach, based on a decomposition into a machine loading and a scheduling sub-problem. The aim of machine loading is to generate a set of efficient (non-dominated) solutions with respect to the load balancing and cost objectives, leaving to the user the task of selecting a compromise solution. Solving the machine loading sub-problem essentially amounts to selecting a process plan for each job and to routing jobs to the machines; then a schedule must be determined. In this paper we deal only with the machine loading sub-problem, as many scheduling methods are already available for the problem with fixed process plans. The machine loading problem is formulated as a bicriterion integer programming model, and two different heuristics are proposed, one based on surrogate duality theory and one based on a genetic descent algorithm. The heuristics are tested on a set of benchmark problems.     

Constraint Propagation and Problem Decomposition: A Preprocessing Procedure for the Job Shop Problem

In recent years, constraint propagation techniques have been shown to be highly effective for solving difficult scheduling problems. In this paper, we present an algorithm which combines constraint propagation with a problem decomposition approach in order to simplify the solution of the job shop scheduling problem. This is mainly guided by the observation that constraint propagation is more effective for small problem instances. Roughly speaking, the algorithm consists of deducing operation sequences that are likely to occur in an optimal solution of the job shop scheduling problem (JSP).The algorithm for which the name edge-guessing procedure has been chosen – since with respect to the job shop scheduling problem (JSP) the deduction of machine sequences is mainly equivalent to orienting edges in a disjunctive graph – can be applied in a preprocessing step, reducing the solution space, thus speeding up the overall solution process. In spite of the heuristic nature of edge-guessing, it still leads to near-optimal solutions. If combined with a heuristic algorithm, we will demonstrate that given the same amount of computation time, the additional application of edge-guessing leads to better solutions. This has been tested on a set of well-known JSP benchmark problem instances.     

On scheduling an unbounded batch machine

A batch machine is a machine that can process up to c jobs simultaneously as a batch, and the processing time of the batch is equal to the longest processing time of the jobs assigned to it. In this paper, we deal with the complexity of scheduling an unbounded batch machine, i.e., c=+∞. We prove that minimizing total tardiness is binary NP-hard, which has been an open problem in the literature. Also, we establish the pseudopolynomial solvability of the unbounded batch machine scheduling problem with job release dates and any regular objective. This is distinct from the bounded batch machine and the classical single machine scheduling problems, most of which with different release dates are unary NP-hard. Combined with the existing results, this paper provides a nearly complete mapping of the complexity of scheduling an unbounded batch machine.     

A neighbourhood scheme with a compressed solution space for the early/tardy scheduling problem

The single machine, distinct due date, early/tardy machine scheduling problem closely models the situation faced by Just-In-Time manufacturers. This paper develops a new method of finding good quality solutions to this scheduling problem by using the concept of a ‘compressed solution space’, based on a binary representation of the early/tardy scheduling problem, and tabu search. A heuristic which simultaneously sequences and schedules the jobs is developed to perform the conversion between the compressed and physical solution spaces. Results show that the compressed solution space performs well with small problems, and is superior to standard tabu search solution spaces for large-scale, realistically sized problems.     

带有退化效应和序列相关运输时间的排序问题

考虑带有退化效应和序列相关运输时间的单机排序问题. 工件的加工时间是其开工时间的简单线性增加函数. 当机器单个加工工件时, 极小化最大完工时间、(加权)总完工时间和总延迟问题被证明是多项式可解的, EDD序对于极小化最大延迟问题不是最优排序, 另外, 就交货期和退化率一致情形给出了一最优算法. 当机器可分批加工工件时, 分别就极小化最大完工时间和加权总完工时间问题提出了多项式时间最优算法.     

基于启发式Q学习的汽车涂装车间作业排序优化

针对汽车涂装车间中的作业优化排序问题,提出一种基于启发式Q学习的优化算法。首先,建立包括满足总装车间生产顺序和最小化喷枪颜色切换次数的多目标整数规划模型。将涂装作业优化排序问题抽象为马尔可夫过程,建立基于启发式Q算法的求解方法。通过具体案例,对比分析了启发式Q学习、Q学习、遗传算法三种方案的优劣。结果表明:在大规模问题域中,启发式Q学习算法具有寻优效率更高、效果更好的优势。本研究为机器学习算法在汽车涂装作业优化排序问题的应用提出了新思路。     

Scheduling arrivals to a production system in a fuzzy environment

A frequently encountered scheduling problem is to determine a material and job ready time while simultaneously finding a production sequence given customer-specified due dates. Often the production times and due dates are vague. This paper presents an investigation of scheduling ready times for a set of jobs with fuzzy service times and due dates. The ready time is constrained in that the possibility that a job is late must not exceed a predefined value. The objective in such an instance is to maximize the ready time without violating these constraints. The steps necessary to determine the maximum ready time and cases in which this effort may be significantly reduced are presented for single machine and flow shop production systems. Finally, a branch and bound technique is developed for cases in which the optimal job sequence cannot be determined a priori.     

Scheduling in a two-machine flowshop with batch processing machine(s) for earliness/tardiness measure under a common due date

This paper considers a scheduling problem for a two-machine flowshop with batch processing machine(s) (BPMs) incorporated where the earliness/tardiness (E/T) measure and a common due date are considered. It assumes that each batch has the same processing time and that the common due date is not set earlier than the total job processing time on the first machine. Under these assumptions, some solution properties are characterized for three different problem cases to derive their associated solution algorithms. For the first two cases concerned with two different machine sequences such as batch-to-discrete and batch-to-batch machine sequences, a polynomial time algorithm is derived based on some of the solution properties. For the last case concerned with discrete-to-batch machine sequence, a pseudopolynomial algorithm is exploited.     

A branch-and-bound algorithm for single-machine scheduling with batch delivery minimizing flow times and delivery costs

This paper addresses scheduling a set of jobs on a single machine for delivery in batches to customers or to other machines for further processing. The problem is a natural extension of minimizing the sum of flow times by considering the possibility of delivering jobs in batches and introducing batch delivery costs. The scheduling objective adopted is that of minimizing the sum of flow times and delivery costs. The extended problem arises in the context of coordination between machine scheduling and a distribution system in a supply chain network. Structural properties of the problem are investigated and used to devise a branch-and-bound solution scheme. Computational experiments show significant improvements over an existing dynamic programming algorithm.     

A Distributed Evolutionary Simulated Annealing Algorithm for Combinatorial Optimisation Problems

In this paper, the Evolutionary Simulated Annealing (ESA) algorithm, its distributed implementation (dESA) and its application to two combinatorial problems are presented. ESA consists of a population, a simulated annealing operator, instead of the more usual reproduction operators used in evolutionary algorithms, and a selection operator. The implementation is based on a multi island (agent) system running on the Distributed Resource Machine (DRM), which is a novel, scalable, distributed virtual machine based on Java technology. As WAN/LAN systems are the most common multi-machine systems, dESA implementation is based on them rather than any other parallel machine. The problems tackled are well-known combinatorial optimisation problems, namely, the classical job-shop scheduling problem and the uncapacitated facility location problem. They are difficult benchmarks, widely used to measure the efficiency of metaheuristics with respect to both the quality of the solutions and the central processing unit (CPU) time spent. Both applications show that dESA solves problems finding either the optimum or a very near optimum solution within a reasonable time outperforming the recent reported approaches for each one allowing the faster solution of existing problems and the solution of larger problems.     

A Two-product,Single-machine,Storage-constrained Production Problem

This paper considers a two-product, single-machine production scheduling problem where there is an added constraint on the amount of finished stock that can be held. The need for a ratio of the cycle times of the two products is a feature both of two-product production scheduling problems contained on one machine and two-product inventory problems with constraints on storage capacity. This means an easy solution algorithm is possible for the problem addressed in the paper which has both types of constraints.     

Single-machine batch scheduling minimizing weighted flow times and delivery costs

This paper addresses scheduling a set of jobs on a single machine for delivery in batches to one customer or to another machine for further processing. The problem is a natural extension of that of minimising the sum of weighted flow times, considering the possibility of delivering jobs in batches and introducing batch delivery costs. The scheduling objective adopted is that of minimising the sum of weighted flow times and delivery costs. The extended problem arises in the context of coordination between machine scheduling and a distribution system in a supply chain network. Structural properties of the problem are investigated and used to devise a branch-and-bound solution method. For the special case, when the maximum number of batches is fixed, the branch-and-bound scheme provided shows significant improvements over an existing dynamic-programming algorithm.     

Single machine stochastic JIT scheduling problem subject to machine breakdowns

In this paper we research the single machine stochastic JIT scheduling problem subject to the machine breakdowns for preemptive-resume and preemptive-repeat.The objective function of the problem is the sum of squared deviations of the job-expected completion times from the due date.For preemptive-resume,we show that the optimal sequence of the SSDE problem is V-shaped with respect to expected processing times.And a dynamic programming algorithm with the pseudopolynomial time complexity is given.We discuss the difference between the SSDE problem and the ESSD problem and show that the optimal solution of the SSDE problem is a good approximate optimal solution of the ESSD problem,and the optimal solution of the SSDE problem is an optimal solution of the ESSD problem under some conditions.For preemptive-repeat,the stochastic JIT scheduling problem has not been solved since the variances of the completion times cannot be computed.We replace the ESSD problem by the SSDE problem.We show that the optimal sequence of the SSDE problem is V-shaped with respect to the expected occupying times.And a dynamic programming algorithm with the pseudopolynomial time complexity is given.A new thought is advanced for the research of the preemptive-repeat stochastic JIT scheduling problem.     

An approach for solving a class of transportation scheduling problems

An algorithm is developed for solving a class of transportation scheduling problems. It applies for a variety of problems such as: the Combining Truck Trip problem, the Delivery problem, the School Bus problem, the Assignment of Buses to Schedules, and the Travelling Salesman problem. The objective functions of the above problems differ from each other. Yet, by using the “savings method” proposed by Clarke and Wright, and extended by Gaskell, we are able to define each one of the above problems as a series of assignment problems. The cost matrix entries of each one of the assignment problems are a function of the constraints of the particular routing or scheduling problem. The solution to the assignment problem determines an upper bound of the optimal solution to the original problem. By combining the above procedure with a Branch and Bound procedure, it is possible to obtain the optimal solution in a finite number of steps. In some cases the Branch and Bound process can be eliminated due to the nature of the problem and in those cases the algorithm is efficient.     

Scheduling flexible machining and assembly systems

Most scheduling papers consider flexible machining and assembly systems as being independent. In this paper, a heuristic two-level scheduling algorithm for a system consisting of a machining and an assembly subsystem is developed. It is shown that the upper level problem is equivalent to the two machine flow shop problem. The algorithm at the lower level schedules jobs according to the established product and part priorities. Related issues, such as batching, due dates, process planning and alternative routes, are discussed. The algorithm and associated concepts are illustrated on a number of numerical examples.     

On non-permutation solutions to some two machine flow shop scheduling problems

In this paper, we study two versions of the two machine flow shop scheduling problem, where schedule length is to be minimized. First, we consider the two machine flow shop with setup, processing, and removal times separated. It is shown that an optimal solution need not be a permutation schedule, and that the problem isNP-hard in the strong sense, which contradicts some known results. The tight worst-case bound for an optimal permutation solution in proportion to a global optimal solution is shown to be 3/2. An O(n) approximation algorithm with this bound is presented. Secondly, we consider the two machine flow shop with finite storage capacity. Again, it is shown that there may not exist an optimal solution that is a permutation schedule, and that the problem isNP-hard in the strong sense.     

A dynamic edge covering and scheduling problem: complexity results and approximation algorithms

We consider a new dynamic edge covering and scheduling problem that focuses on assigning resources to nodes in a network to minimize the amount of time required to process all edges in it. Resources need to be co-located at the endpoints of an edge for it to be processed and, therefore, this problem contains both edge covering and scheduling decisions. These new problems have motivating applications in traffic systems and military intelligence operations. We provide complexity results for the dynamic edge covering and scheduling problem over different types of networks. We then show that existing approximation algorithms for parallel machine scheduling problems can be leveraged to provide approximation algorithms for this new class of problems over certain types of networks.     

Machine scheduling with an availability constraint

Most literature in scheduling assumes that machines are available simultaneously at all times. However, this availability may not be true in real industry settings. In this paper, we assume that the machine may not always be available. This happens often in the industry due to a machine breakdown (stochastic) or preventive maintenance (deterministic) during the scheduling period. We study the scheduling problem under this general situation and for the deterministic case.We discuss various performance measures and various machine environments. In each case, we either provide a polynomial optimal algorithm to solve the problem, or prove that the problem is NP-hard. In the latter case, we develop pseudo-polynomial dynamic programming models to solve the problem optimally and/or provide heuristics with an error bound analysis.This research was supported in part by NSF grant DDM 9201627     

Evolutionary hybrid approaches for generation scheduling in power systems

Generation scheduling (GS) in power systems is a tough optimisation problem which continues to present a challenge for efficient solution techniques. The solution is to define on/off decisions and generation levels for each electricity generator of a power system for each scheduling interval. The solution procedure requires simultaneous consideration of binary decision and continuous variables. In recent years researchers have focused much attention on developing new hybrid approaches using evolutionary and traditional exact methods for this type of mixed-integer problems. This paper investigates how the optimum or near optimum solution for the GS problem may be quickly identified. A design is proposed which uses a variety of metaheuristic, heuristics and mathematical programming techniques within a hybrid framework. The results obtained for two case studies are promising and show that the hybrid approach offers an effective alternative for solving the GS problems within a realistic timeframe.