A FRAMEWORK FOR MACHINE SCHEDULING PROBLEMS WITH CONTROLLABLE PROCESSING TIMES

This is a study of single and parallel machine scheduling problems with controllable processing time for each job. The processing time for job j depends on the position of the job in the schedule and is a function of the number of resource units allocated to its processing. Processing time functions and processing cost functions are allowed to be nonlinear. The scheduling problems considered here have important applications in industry and include many of the existing scheduling models as special cases. For the single machine problem, the objective is minimization of total compression costs plus a scheduling measure. The scheduling measures include makespan, total flow time, total differences in completion times, total differences in waiting times, and total earliness and tardiness with a common due date for all jobs. Except when the total earliness and tardiness measure is involved, each case the problem is solved efficiently. Under an assumption typically satisfied in just-in-time systems, the problem with total earliness and tardiness measure is also solved efficiently. Finally, for a large class of processing time functions; parallel machine problems with total flow time and total earliness and tardiness measures are solved efficiently. In each case we reduce the problem to a transportation problem.     

Multi-objective scheduling of two-job families on a single machine

This paper presents a preliminary analysis of the typical scheduling environment in semiconductor manufacturing involving multiple job families, and where more than one objective such as cycle time, machine utilization and the due-date accuracy needs to be simultaneously considered. In this study, the NP-hard problem of scheduling N independent jobs on a single testing machine with due dates and sequence-dependent setup times is addressed, where the multiple objectives are to minimize average cycle time, to minimize average tardiness, and to maximize machine utilization. A Pareto optimal solution, which is not inferior to any other feasible solutions in terms of all objectives, is generated combining the analytically optimal and conjunctive simulated scheduling approach. First, the machine-scheduling problem is modeled using the discrete event simulation approach and the problem is divided into simulation clock based lot selection sub-problems. Then, a Pareto optimal lot is selected using the compromise programming technique for multiobjective optimization at each decision instant in simulated time. With the help of a broad experimental design, this developed solution is then compared with common heuristic-dispatching rules such as SPT and EDD, which show better results for all the objectives over a wide range of problems. The developed scheduling method shows approximately 16.7% reduction in average cycle time, 25.6% reduction in average tardiness, and 21.6% improvement in machine utilization over the common dispatching rules, SPT and EDD.     

Minmax common flow-allowance problems with convex resource allocation and position-dependent workloads

We study minmax due-date based on common flow-allowance assignment and scheduling problems on a single machine, and extend known results in scheduling theory by considering convex resource allocation. The total cost function of a given job consists of its earliness, tardiness and flow-allowance cost components. Thus, the common flow-allowance and the actual jobs’ processing times are decision variables, implying that the due-dates and actual processing times can be controlled by allocating additional resource to the job operations. Consequently, our goal is to optimize a cost function by seeking the optimal job sequence, the optimal job-dependent due-dates along with the actual processing times. In all addressed problems we aim to minimize the maximal cost among all the jobs subject to a constraint on the resource consumption. We start by analyzing and solving the problem with position-independent workloads and then proceed to position-dependent workloads. Finally, the results are generalized to the method of common due-window. For all studied problems closed form solutions are provided, leading to polynomial time solutions.

     

A hybrid approach to sequencing jobs using heuristic rules and neural networks

A hybrid approach to solve job sequencing problems using heuristic rules and artificial neural networks is proposed. The problem is to find a job sequence for a single machine that minimizes the total weighted tardiness of the jobs. Two different cases are considered: (1) when there are no setups, and (2) when there are sequence-dependent setup times. So far, successful heuristic rules for these cases are: apparent tardiness cost (ATC) rule proposed by Vepsalainen and Morton for the former case, and an extended version of the ATC rule (ATCS) proposed by Lee, Bhaskaran, and Pinedo for the latter. Both approaches utilize some look-ahead parameters for calculating the priority index of each job. As reported by Bhaskaran and Pinedo, the proper value of the look-ahead parameter depends upon certain problem characteristics, such as due-date tightness and due-date range. Thus, an obvious extension of the ATC or the ATCS rule is to adjust the parameter values depending upon the problem characteristics: this is known to be a difficult task. In this paper, we propose an application of a neural network as a tool to ‘predict’ proper values of the look-ahead parameters. Our computational tests show that the proposed hybrid approach outperforms both the ATC rule with a fixed parameter value and the ATCS using the heuristic curve-fitting method.     

The no-wait job shop with regular objective: a method based on optimal job insertion

The no-wait job shop problem (NWJS-R) considered here is a version of the job shop scheduling problem where, for any two operations of a job, a fixed time lag between their starting times is prescribed. Also, sequence-dependent set-up times between consecutive operations on a machine can be present. The problem consists in finding a schedule that minimizes a general regular objective function. We study the so-called optimal job insertion problem in the NWJS-R and prove that this problem is solvable in polynomial time by a very efficient algorithm, generalizing a result we obtained in the case of a makespan objective. We then propose a large neighborhood local search method for the NWJS-R based on the optimal job insertion algorithm and present extensive numerical results that compare favorably with current benchmarks when available.     

Single facility scheduling with dual criteria: minimizing maximum tardiness subject to minimum number of tardy jobs

This paper considers the static single machine scheduling problem with the objective of minimizing the maximum tardiness of any job subject to the constraint that the total number of lardy jobs is minimum. Based on simple dominance conditions an o(n²) heuristic algorithm is proposed to find an approximate solution to this problem. The effectiveness of the proposed heuristic algorithm is empirically evaluated by solving a large number of problems and comparing them to the optimal solutions obtained through the branch and bound algorithm.     

Optimal Policy for a Stochastic Scheduling Problem with Applications to Surgical Scheduling

We consider the stochastic, single‐machine earliness/tardiness problem (SET), with the sequence of processing of the jobs and their due‐dates as decisions and the objective of minimizing the sum of the expected earliness and tardiness costs over all the jobs. In a recent paper, Baker ( 2014 ) shows the optimality of the Shortest‐Variance‐First (SVF) rule under the following two assumptions: (a) The processing duration of each job follows a normal distribution. (b) The earliness and tardiness cost parameters are the same for all the jobs. In this study, we consider problem SET under assumption (b). We generalize Baker's result by establishing the optimality of the SVF rule for more general distributions of the processing durations and a more general objective function. Specifically, we show that the SVF rule is optimal under the assumption of dilation ordering of the processing durations. Since convex ordering implies dilation ordering (under finite means), the SVF sequence is also optimal under convex ordering of the processing durations. We also study the effect of variability of the processing durations of the jobs on the optimal cost. An application of problem SET in surgical scheduling is discussed.     

Heterogeneous dispatching rules in job and flow shops

This paper reports the results of a study of the use of heterogeneous dispatching rules for the scheduling of work in a job shop. The methodology employed included discrete event simulation, using rule combinations determined by prior genetic algorithm searches and generalization using neural networks. Eight dispatching rules were considered, including first in first out (FIFO), earliest due date ( EDD), shortest processing time (SPT), slack/ number of operations (SLK), critical ratio (CR), modified due date (MDD), modified operation due date (MOD), and apparent tardiness cost (ATC). A three-machine job shop was studied, in which three work organizations were employed, pure flow (fixed sequence), pure job shop ( random sequence), and a hybrid shop where flow is random but with unequal probabilities. Three levels of machine loading were used and average tardiness was used as the performance measure. In most cases, modified due date and apparent tardiness cost were the best rules. The application of the best rules effected the results primarily when applied to bottleneck machines or the first machine in a pure flow shop. Nearly any other rule was acceptable on non-botdeneck machines except FIFO and CR, which consistently perform poorly. No major advantage of mixing rules was found.     

Optimal job insertion in the no-wait job shop

The no-wait job shop (NWJS) considered here is a version of the job shop scheduling problem where, for any two operations of a job, a fixed time lag between their starting times is given. Also, sequence-dependent set-up times between consecutive operations on a machine can be present. The NWJS problem consists in finding a schedule that minimizes the makespan. We address here the so-called optimal job insertion problem (OJI) in the NWJS. While the OJI is NP-hard in the classical job shop, it was shown by Gröflin & Klinkert to be solvable in polynomial time in the NWJS. We present a highly efficient algorithm with running time $\mathcal {O}(n^{2}\cdot\max\{n,m\})$ for this problem. The algorithm is based on a compact formulation of the NWJS problem and a characterization of all feasible insertions as the stable sets (of prescribed cardinality) in a derived comparability graph. As an application of our algorithm, we propose a heuristic for the NWJS problem based on optimal job insertion and present numerical results that compare favorably with current benchmarks.     

An integer programming approach to scheduling in a contaminated area

This paper presents a time-indexed integer programming formulation for scheduling dependent jobs executed by a team of workers in an area contaminated with radio-active or chemical materials. The dynamics of the harmful factor and the norms of organism recovery imply that each work period for a job should be immediately followed by a rest period for the worker executing this job and the length of the rest period depends on the start time of the corresponding work period. The problem is modeled as an NP-hard problem of scheduling on unrelated parallel processors with start time dependent processing times and different objective functions: maximum or total completion time and maximum or total tardiness. The special case of scheduling jobs executed by a single worker is also considered. Numerical examples and some computational results are reported.     

A tardiness-augmented approximation scheme for rejection-allowed multiprocessor rescheduling

In the multiprocessor scheduling problem to minimize the total job completion time, an optimal schedule can be obtained by the shortest processing time rule and the completion time of each job in the schedule can be used as a guarantee for scheduling revenue. However, in practice, some jobs will not arrive at the beginning of the schedule but are delayed and their delayed arrival times are given to the decision-maker for possible rescheduling. The decision-maker can choose to reject some jobs in order to minimize the total operational cost that includes three cost components: the total rejection cost of the rejected jobs, the total completion time of the accepted jobs, and the penalty on the maximum tardiness for the accepted jobs, for which their completion times in the planned schedule are their virtual due dates. This novel rescheduling problem generalizes several classic NP-hard scheduling problems. We first design a pseudo-polynomial time dynamic programming exact algorithm and then, when the tardiness can be unbounded, we develop it into a fully polynomial time approximation scheme. The dynamic programming exact algorithm has a space complexity too high for truthful implementation; we propose an alternative to integrate the enumeration and the dynamic programming recurrences, followed by a depth-first-search walk in the reschedule space. We implemented the alternative exact algorithm in C and conducted numerical experiments to demonstrate its promising performance.

     

MINIMIZING ABSOLUTE AND SQUARED DEVIATION OF COMPLETION TIMES FROM DUE DATES

This is a study of a single-machine scheduling problem with the objective of minimizing the sum of a function of earliness and tardiness called the earliness and tardiness (ET) problem. I will show that if priority weights of jobs are proportional to their processing times, and if earliness and tardiness cost functions are linear, the problem will be equivalent to the total weighted tardiness problem. This proves that the et problem is np -hard. In addition, I present a heuristic algorithm with worst case bound for the et problem based on the equivalence relation between the two. When earliness and tardiness cost functions are quadratic, I consider the problem for a common due date for all jobs and for different job due dates. In general, the et problem with quadratic earliness and tardiness cost functions and all job weights equal to one is np -hard. I show that in many cases, when weights of jobs are proportional to their processing times, the problem can be solved efficiently. In the published results on the et problem with quadratic earliness and tardiness cost functions other researchers have assumed a zero starting time for the schedule. I discuss the advantages of a nonzero starting time for the schedule.     

Resource constrained scheduling with general truncated job-dependent learning effect

Scheduling with general truncated job-dependent learning effect and resource-dependent processing times is studied on a single machine. It is assumed that the job processing time is a function of the amount of resource allocated to the job, the general job-dependent learning effect and the job-dependent control parameter. For each version of the problem that differs in terms of the objective functions and the processing time functions, the optimal resource allocation is provided. Polynomial time algorithms are also developed to find the optimal schedule of several versions of the problem.     

MINIMIZING TOTAL WEIGHTED COMPLETION TIME ON A SINGLE BATCH PROCESSING MACHINE

We study the problem of scheduling jobs on a single batch processing machine to minimize the total weighted completion time. A batch processing machine is one that can process a number of jobs simultaneously as a batch. The processing time of a batch is given by the processing time of the longest job in the batch. We present a branch and bound algorithm to obtain optimal solutions and develop lower bounds and dominance conditions. We also develop a number of heuristics and evaluate their performance through extensive computational experiments. Results show that two of the heuristics consistently generate high-quality solutions in modest CPU times.     

混合离散差分进化算法在单机批处理调度中的应用

本文研究单机批处理调度问题,批处理机有批次容量限制,批处理时间由每个批次所含作业中的最长作业处理时间决定。每个作业具有不同的大小、处理时间、提前拖期惩罚权重,所有作业具有公共交货期,且交货期无限晚。目标函数为最小化所有作业的加权提前拖期惩罚之和。该问题已被证明为NP难题,本研究找到了其最优解具有的一些性质,在此基础上利用它们提出了一种动态规划(DP)与差分进化(DE)算法相结合的混合离散差分进化(HDDE)算法来求解该问题,通过与传统的遗传算法、模拟退火算法和迭代贪婪算法进行对比,HDDE算法显示了更加强大的全局搜索能力。     

Total tardiness minimization on unrelated parallel machine scheduling with auxiliary equipment constraints

This research deals with scheduling jobs on unrelated parallel machines with auxiliary equipment constraints. Each job has a due date and requires a single operation. A setup for dies is incurred if there is a switch from processing one type of job to another type. For a die type, the number of dies is limited. Due to the attributes of the machines and the fitness of dies to each, the processing time for a job depends on the machine on which the job is processed, each job being restricted to processing on certain machines. In this paper, an effective heuristic based on threshold-accepting methods, tabu lists, and improvement procedures is proposed to minimize total tardiness. An extensive experiment is conducted to evaluate the computational characteristics of the proposed heuristic. Computational experiences demonstrate that the proposed heuristic is capable of obtaining optimal solutions for small-sized problems, and significantly outperforms an ATCS procedure and a simulated annealing method for problems in larger sizes.     

Online scheduling with rejection and reordering: exact algorithms for unit size jobs

We study an online scheduling problem with rejection on \(m\ge 2\) identical machines, in which we deal with unit size jobs. Each arriving job has a rejection value (a rejection cost or penalty for minimization problems, and a rejection profit for maximization problems) associated with it. A buffer of size \(K\) is available to store \(K\) jobs. A job which is not stored in the buffer must be either assigned to a machine or rejected. Upon the arrival of a new job, the job can be stored in the buffer if there is a free slot (possibly created by evicting other jobs and assigning or rejecting every evicted job). At termination, the buffer must be emptied. We study four variants of the problem, as follows. We study the makespan minimization problem, where the goal is to minimize the sum of the makespan and the penalty of rejected jobs, and the \(\ell _p\) norm minimization problem, where the goal is to minimize the sum of the \(\ell _p\) norm of the vector of machine completion times and the penalty of rejected jobs. We also study two maximization problems, where the goal in the first version is to maximize the sum of the minimum machine load (the cover value of the machines) and the total rejection profit, and in the second version the goal is to maximize a function of the machine completion times (which measures the balance of machine loads) and the total rejection profit. We show that an optimal solution (an exact solution for the offline problem) can always be obtained in this environment, and determine the required buffer size. Specifically, for all four variants we present optimal algorithms with \(K=m-1\) and prove that in each case, using a buffer of size at most \(m-2\) does not allow the design of an optimal algorithm, which makes our algorithms optimal in this respect as well. The lower bounds hold even for the special case where the rejection value is equal for all input jobs.     

Single-machine scheduling with periodic due dates to minimize the total earliness and tardy penalty

We consider a single-machine scheduling problem such that the due dates are assigned to each job depending on its order, and the lengths of the intervals between consecutive due dates are identical. The objective is to minimize the total penalty for the earliness and tardiness of each job. The early penalty proportionally increases according to the earliness amount, while the tardy penalty increases according to the step function. We show that the problem is strongly NP-hard, and furthermore, polynomially solvable if the two types of processing times exist.

     

Minimizing tardiness on a single processor with sequence-dependent setup times: a simulated annealing approach

In today's fast-paced Just-In-Time and mass customization manufacturing in a sequence-dependent setup environment, the challenge of making production schedules to meet due-date requirements is becoming a more complex problem. Unfortunately, much of the research on operations scheduling problems has either ignored setup times or assumed that setup times on each machine are independent of the job sequence. This paper considers the problem of minimizing tardiness, a common measure of due-date performance, in a sequence-dependent setup environment. Simulated annealing was used to solve the sequencing problem, and its performance was compared with random search. Our experimental results show that the algorithm can find a good solution fairly quickly, and thus can rework schedules frequently to react to variations in the schedule. The algorithm is invaluable for ‘on-line’ production scheduling and ‘last-minute’ changes to production schedule. The results of this research also suggest ways in which more complex and realistic job shop environments, such as multiple machines with a higher number of jobs in the sequence, and other scheduling objectives can be modeled. This research also investigates computational aspects of simulated annealing in solving complex scheduling problems.     

Sequential Search and Selection Problem Under Uncertainty

This paper formulates and discusses a series of sequential decision problems of the following common structure: A decision alternative of multiple attributes‐that is, a job, an employee, or an investment alternative‐is to be selected within a certain fixed length of time. An unknown number of alternatives are presented sequentially, either deterministically or in a random manner. The decision maker can rank all the alternatives from best to worst without ties, and the decision to accept or reject an alternative is based solely on the relative ranks of those alternatives evaluated so far. The nonparametric sequential decision problem is first studied for a model involving a discrete time period and then generalized in terms of continuous time. Also considered is a variant of this problem involving a Bayesian estimation of (1) the uncertain probability of having an alternative at a given stage in the discrete‐time model and (2) the arrival rate of alternatives in the continuous‐time model. The optimal selection strategy that maximizes the probability of selecting the absolute best alternative is illustrated with the job search problem and the single‐machine job assignment problem.