Genetic algorithms for single machine scheduling with time-dependent deterioration and rate-modifying activities
Highlights
► We consider single machine scheduling with time-dependent deterioration and RMAs. ► We derive a mathematical model to minimize the makespan for the scheduling problem. ► We propose genetic algorithms (GAs) with three different chromosome types. ► We evaluate the performances of the GAs using several randomly generated examples.
Introduction
In the classical scheduling, the common assumption is that processing times of the jobs are fixed and known a priori with certainty. In practice, however, machine production rate becomes less than normal due to a mal-position of tools, mal-alignment of jobs, abrasion of tools, and scraps of operations, etc. In these situations, the processing time of the jobs increases depending upon the sequence of jobs or the starting time of jobs. Such problems are generally known as machine scheduling problems with deterioration. The deteriorated processing time of jobs is recovered to an original processing time by the maintenance or cleaning process of machines. The recovering process bringing back to a normal processing time is called rate-modifying activity (RMA) (Lee & Leon, 2001).
The phenomenon known as deterioration has been received increasing attention over the last two decades in recent years, after pioneering efforts of scheduling models by Gupta and Gupta (1988) and Browne and Yechiali (1990). They considered a single machine scheduling problem with deteriorating jobs. Gupta and Gupta (1988) introduced a single machine scheduling problem with the processing time of a task being a polynomial function of its starting time. Browne and Yechiali (1990) considered a single machine scheduling problem with the job processing time of a task being a non-decreasing with start time-dependent linear function. Alidaee and Womer (1999) reviewed a comprehensive review of job processing time is time-dependent and provide various forms of the function of processing time including linear, piecewise linear, and nonlinear. Their work demonstrated the majority of the literatures for the processing time function with single machine environment. Cheng, Ding, and Lin (2004) also surveyed a class of machine scheduling problems in which processing time of a task is dependent on its starting time in a schedule. They classified scheduling problems with various time-dependent deterioration functions and provided a framework to illustrate how models for the class of problems have been generalized from the classical scheduling theory. Cheng and Ding (2000) proposed a single machine scheduling model to minimize makespan with deadlines and increasing rates of processing times. They found that both problems are solvable by a dynamic programming algorithm. Bachman, Janiak, and Kovalyov (2002) proved the NP-hardness of total weighted completion time for single machine scheduling in which the job processing times are decreasing linear functions dependent on their start times. Wang, Sun, and Sun (2011) studied single-machine total completion time scheduling with a time-dependent deterioration. They proposed two heuristic algorithms utilizing the V-shaped property. Since the complexity of the problem remains open, they compared the worst case error bound using SPT with two proposed heuristics. Several research focus on developing meta-heuristics. Shin, Kim, and Kim (2002) presented a heuristic and tabu search algorithm for the single-machine scheduling problem with sequence-dependent setup times, release times, and due dates, with the objective function of minimizing the maximum lateness. Stecco, Cordeau, and Moretti (2009) developed a tabu search heuristic for a sequence-dependent and time-dependent scheduling problem on a single machine. Bahalke, Yolmeh, and Shahanaghi (2010) proposed several hybrid meta-heuristic algorithms for the single-machine scheduling problem with sequence dependent setup time and deteriorating jobs, with minimize makespan.
Meanwhile, in RMA, Lee and Leon (2001) initially introduced a single machine scheduling with a RMA. They assumed that one rate-modifying activity which changes the production rate of equipment was included during the planning horizon. Given a set of jobs to be performed by a resource and an RMA of fixed length that recovers the processing rate of the resource, they determined (i) the sequence in which the jobs should be performed and (ii) when to schedule the fixed length RMA so that the objective of the scheduling is optimized. They separately classified the scheduling problem to following four objective functions: makespan, total completion time, total weighted completion time, and maximum lateness and each of the cases propose a dynamic programming algorithm after they found optimality properties. Qi, Chen, and Tu (1999) considered a single machine problem with the possibility for multiple maintenance activities, but during scheduling period they ignored the deterioration of processing time for jobs. Lee and Lin (2001) considered single machine scheduling problems involving repair and maintenance activities. They derived optimal policies for scheduling fixed length RMAs and job sequencing in an environment by machine breakdowns. Mosheiov and Sarig (2009) analyzed a single machine scheduling determining both sequencing of jobs and positioning of a RMA with common due-window assignment. Lee and Chen (2000) studied a parallel scheduling problem assuming that single RMA was processed for each machine in entire planning horizon. Zhao, Tang, and Cheng (2009) presented two parallel machines to minimize the makespan assuming that single RMA was processed for each machine. Iranpoor, Ghomi, and Zandieh (forthcoming) presented a single machine scheduling problem with sequence-dependent setup times and single RMA.
A review of the above literature reveals that most of studies only focused on machine scheduling problems with deterioration or machine scheduling problem with RMAs, even if the deterioration of jobs and RMAs are interrelated with each other. Lodree and Geiger (2010) studied scheduling problems integrated with time-dependent processing times and single RMA. Under a certain condition, they provided the optimal policy determining optimal position of RMA and sequencing of jobs. Zhao and Tang (2012) considered single machine scheduling problem with positioning of single RMA and deterioration of jobs. Lodree and Geiger (2010) and Zhao and Tang (2012) defined RMA as an activity changing the normal processing times of the jobs following this activity, and assumed that the processing time of a job is a linear function of its starting time and the job-independent deterioration rates are identical for all jobs. Kim and Joo (2011) studied single machine scheduling model with multiple RMAs and the deterioration of jobs nonlinearly increased by the difference between the position of the job and the previous RMA. In their study, the RMAs are defined as a recovering process bringing back to a normal processing time.
In this study, single machine scheduling problem with time-dependent deterioration and multiple RMAs are considered. The RMAs are defined as a recovering process bringing back to a normal processing time like Lee and Leon (2001) and Kim and Joo (2011). There are N independent jobs to be scheduled. Job j has original processing time pj and time-dependent deterioration rate rj. The actual processing time of job j is increased (deteriorated) depending upon the gap xj between recent RMA and starting time of the job, hence the time becomes . The deteriorated processing time of jobs is recovered to an original processing time by RMAs. The RMA time Q is assumed to be fixed. The working horizons between RMAs including horizons before the first and after the last RMAs are called buckets in this paper. The goal of the problem is to find a schedule minimizing the makespan which is the sum of each completion time Ck of all the assigned jobs in k th bucket and all RMA times. The schedule includes the assigning jobs to buckets and the sequencing of assigned jobs in each bucket, along with the positions of RMAs. Fig. 1 describes the scheduling of jobs in the example.
The paper organized as follows. Section 2 presents the notations and a mathematical model finding the optimal solution for the problem. Three genetic algorithms (GAs) are proposed with different chromosome representations in Section 3. In Section 4, the performances of the GAs are evaluated through computational experiments. Finally, summary and further research areas are reviewed in Section 5.
Section snippets
Model formulation
If the buckets are regarded as parallel machines, the scheduling problem with time-dependent deterioration and rate-modifying activities are similar to the parallel machine scheduling problem with time-dependent deterioration. So, the mathematical model of the problem is formulated with the similar structure of parallel machine scheduling model. In this section, the mixed integer programming model is derive to minimize the makespan. The job assignments to buckets and the sequences of assigned
Genetic algorithms
We propose an effective and efficient meta-heuristic algorithm, because the mixed integer programming model is not suitable for the problems over 10 jobs because of the long computation time (See Section 4). The single machine scheduling problem with time-dependent deterioration and multiple RMAs considered in this paper is typical combinatorial optimization problem, and GA is known as an effective and efficient algorithm for combinatorial optimization problems (Balin, 2011, Chan et al., 2011,
Computational results
In this section, extensive computational experiments are conducted to evaluate the performance of GA_SC, GA_DS, and GA_DR. Two problem groups are randomly generated according to the total number of jobs because the complexity of the problem highly depends on the number of jobs. Group 1, which involves problems with less than 10 jobs, is for comparing solutions obtained by GAs with the optimal solution. To obtain the optimal solution through mathematical programming presented in Section 2, ILOG
Conclusions
In this paper, problems dealing with the integration of two emerging classes of scheduling problems, which are the class of scheduling problems with time-dependent processing times and the class of scheduling problems with rate-modifying activities (RMAs) are examined.
The objective of the study is to find a schedule minimizing the makespan. The schedule includes the assigning jobs to buckets and the sequencing of assigned jobs in each bucket, along with the positions of RMAs. A mixed integer
References (25)
- et al.
Minimizing the total weighted completion time of deteriorating jobs
Information Processing Letters
(2002) Non-identical parallel machine scheduling using genetic algorithm
Expert Systems with Applications
(2011)- et al.
A genetic algorithm-based scheduler for multiproduct parallel machine sheet metal job shop
Expert Systems with Applications
(2011) - et al.
A concise survey of scheduling with time-dependent processing times
European Journal of Operational Research
(2004) - et al.
Single facility scheduling with nonlinear processing times
Computers and Industrial Engineering
(1988) - et al.
Single-machine scheduling with maintenance and repair rate-modifying activities
European Journal of Operational Research
(2001) - et al.
Machine scheduling with a rate-modifying activity
European Journal of Operational Research
(2001) - et al.
A note on the optimal sequence position for a rate-modifying activity under simple linear deterioration
European Journal of Operational Research
(2010) - et al.
Scheduling a maintenance activity and due-window assignment on a single machine”
Computers & Operations Research
(2009) - et al.
Single-machine total completion time scheduling with a time-dependent deterioration
Applied Mathematical Modelling
(2011)
A note to due-window assignment and single machine scheduling with deteriorating jobs and a rate-modifying activity
Computers & Operations Research
Two-parallel machines scheduling with rate-modifying activities to minimize total completion time
European Journal of Operational Research
Cited by (27)
Comparison study of dispatching rules and heuristics for online scheduling of single machine scheduling problem with predicted release time jobs
2024, Expert Systems with ApplicationsAdaptive genetic algorithm for two-stage hybrid flow-shop scheduling with sequence-independent setup time and no-interruption requirement
2022, Expert Systems with ApplicationsCitation Excerpt :If α/g ≥ β, due to the setup request for M1i, i∈ℕg, at Stage 1, the productivity at Stage 1 is always lower than that at Stage 2, leading to an unbalanced manufacturing system and some machines at Stage 2 would stop working from time to time. In this case, the scheduling problem of the system is similar to scheduling a single-machine and a lot of efforts have been done for such a problem (Soltani et al., 2010; Khorshidian et al., 2011; Joo & Kim, 2013; Yin et al., 2014; Perez-Gonzalez & Framinan, 2018; Wan & Yuan, 2018; Goldengorin & Romanuke, 2021). Thus, to increase the productivity, the system is designed such that α/g < β.
Fatigue, personnel scheduling and operations: Review and research opportunities
2021, European Journal of Operational ResearchCitation Excerpt :The deteriorating job scheduling problem is introduced by Browne and Yechiali (1990), who consider a single machine scheduling problem with deteriorating jobs. Since then, machine scheduling problems with deteriorating jobs have received more attention (Joo & Kim, 2013; Sun & Geng, 2019); however, this literature requires more empirical support than it presently enjoys to enable it to mitigate fatigue in practical situations. The literature of occupational health and safety defines fatigue as both a decrement in performance and a subjective feeling (Mountstephen & Sharpe, 1997), which is consistent with our definition in Section 1.
Matheuristic approaches for parallel machine scheduling problem with time-dependent deterioration and multiple rate-modifying activities
2018, Computers and Operations ResearchCitation Excerpt :Subsequently, several studies focused on various deteriorating strategies and multiple RMAs for single or parallel machine scheduling problems. With respect to single machine scheduling problems, Joo and Kim (2013) investigated a machine scheduling problem with deterioration under the environment of the constrained number of RMAs. They provided a novel mathematical model and proposed a genetic algorithm for large sized problems.
A rule-based genetic algorithm with an improvement heuristic for unrelated parallel machine scheduling problem with time-dependent deterioration and multiple rate-modifying activities
2017, Computers and Industrial EngineeringCitation Excerpt :This problem simultaneously considers deteriorating processing times of jobs and RMAs as maintenance activities. The phenomenon that deteriorates (or increases) processing times is known as the deterioration effect (Joo & Kim, 2013). In real manufacturing industry situations, the deterioration effect lets the completion times of jobs increase since there are many factors that affect productivity such as mal-positions of tools, mal-alignments of jobs, and scraps of operations.