Dipak Laha

An improved heuristic for permutation flowshop scheduling

Flowshop scheduling deals with determination of optimum sequence of jobs to be processed on some machines in a fixed order so as to satisfy certain scheduling criteria. The general problem of scheduling has been shown to be NP-complete. Exact algorithms, such as integer programming and branch-and-bound, guarantee optimality but do not yield the optimum solution in polynomial time even for problems of small size. Heuristics have been shown to yield good working solutions (not necessarily optimal) in reasonable time. Although much research on the flowshop problem has been done over several decades starting from Johnsons lgorithm, only a few good algorithms exist. The Nawaz-Enscore-Ham heuristic, used for minimisation of makespan, continues to be the most popular algorithm because of its simplicity, solution quality and time-complexity. In the present paper we have modified the NEH algorithm, achieving significant improvement in the quality of the solution while maintaining the same algorithmic complexity. The proposed approach derives its strength from the use of a population-based technique. Experimental comparisons have been made on a large number of randomly generated test problems of varying problem sizes. Our approach is shown to outperform both the original NEH and NEHs best-known competitor to date, the HFC heuristic. Statistical tests of significance are performed to substantiate the claims of improvement.

An efficient stochastic hybrid heuristic for flowshop scheduling

The paper addresses the problem of flowshop scheduling in order to minimize the makespan objective. Three probabilistic hybrid heuristics are presented for solving permutation flowshop scheduling problem. The proposed methodology combines elements from both constructive heuristic search and a stochastic improvement technique. The stochastic method used in this paper is simulated annealing (SA). Experiments have been run on a large number of randomly generated test problems of varying jobs and machine sizes. Our approach is shown to outperform best-known existing heuristics, including the classical NEH technique (OMEGA, 1983) and the SA based on (OMEGA, 1989) of Osman and Potts . Statistical tests of significance are performed to substantiate the claims of improvement.

Modeling of steelmaking process with effective machine learning techniques

Machine learning methods on open-hearth steel making process prediction.Predicted yield based on 4 machine learning methods has very low errors.Best performance is achieved by SVR. Monitoring and control of the output yield of steel in a steelmaking shop plays a critical role in steel industry. The yield of steel determines how much percentage of hot metal, scrap, and iron ore are being converted into steel ingots. It represents the operational efficiency of the steelmaking shop and is considered as an important performance measure for producing a specific quantity of steel. Due to complexity of the steelmaking process and nonlinear relationship between the process parameters, modeling the input-output process parameters and accurately predicting the output yield in the steelmaking shop is very difficult and has been a major research issue. Statistical models and artificial neural networks (ANN) have been extensively studied by researchers and practitioners to model a variety of complex processes. In the present study, we consider random forests (RF), ANN, dynamic evolving neuro-fuzzy inference system (DENFIS) and support vector regression (SVR) as competitive learning tools to verify the suitability of applications of these approaches and investigate their comparative predictive ability. In the present investigation, 0.00001 of MSE is set as a goal of learning during modeling. Based on real-life data, the computational results depict that the training and testing MSE values of SVR and DENFIS are close to 0.00001 indicating that they have higher prediction ability than ANN and RF. Also, mean absolute percentage prediction errors of the proposed models confirm that the predicted yield based on each method is in good agreement with the testing datasets. Overall, SVR performs best and DENFIS the next best followed by ANN and RF methods respectively. The results suggest that the prediction precision given by SVR can meet the requirement for the actual production of steel.

A new cuckoo search algorithm for 2-machine robotic cell scheduling problem with sequence-dependent setup times

Abstract The paper addresses the problem of 2-machine robotic cell scheduling of one-unit cycle with sequence-dependent setup times and different loading/unloading times of the parts. As an alternative metaheuristic algorithm, the cuckoo search algorithm has recently attracted growing interests of researchers. It has the capability to search globally as well as locally to converge to the global optimality by exploring the search space more efficiently due to its global random walk governed by Levy flights, rather than standard isotropic random walk. In this study, a discrete cuckoo search algorithm is proposed to determine the sequence of robot moves along with the sequence of parts so that the cycle time is minimized. In the proposed algorithm, the fractional scaling factor based procedure is presented to determine the step length of Levy flights distribution in discrete from and then, using this step length, two neighborhood search techniques, interchange and cyclical shift methods are applied to the current solution to obtain improved solution. A response surface methodology based on desirability function is used to enhance the convergence speed of the proposed algorithm. Also, a design of experiment is employed to tune the operating parameters of the algorithm. Finally, empirical results with a large number of randomly generated problem instances involving large part sizes varying from 200 to 500 under different operating conditions are compared with two well-known algorithms in the literature and demonstrate the effectiveness of the proposed algorithm.

An improved heuristic to minimize total flow time for scheduling in the m-machine no-wait flow shop

In this paper, we present a constructive heuristic to minimize total flow time criterion for the well-known NP-hard no-wait flow shop scheduling problem. It is based on the assumption that the priority of a job in the initial sequence is given by the sum of its processing times on the bottleneck machines. The initial sequence of jobs thus generated is further improved using a new job insertion technique. We show, through computational experimentation, that the proposed method significantly outperforms the best-known heuristics while retaining its time complexity of O(n^2). Statistical tests of significance are used to confirm the improvement in solution quality.

A penalty-shift-insertion-based algorithm to minimize total flow time in no-wait flow shops

This paper proposes a penalty-shift-insertion (PSI)-based algorithm for the no-wait flow shop scheduling problem to minimize total flow time. In the first phase, a penalty-based heuristic, derived from Vogel’s approximation method used for the classic transportation problem is used to generate an initial schedule. In the second phase, a known solution is improved using a forward shift heuristic. Then the third phase improves this solution using a job-pair and a single-job insertion heuristic. Results of the computational experiments with a large number of randomly generated problem instances show that the proposed PSI algorithm is relatively more effective and efficient in minimizing total flow time in a no-wait flow shop than the state-of-the-art procedures. Statistical significance of better results obtained by the proposed algorithm is also reported.

A Hungarian penalty-based construction algorithm to minimize makespan and total flow time in no-wait flow shops

We solve the no-wait flowshop problems with makespan and mean flowtime objectives.For the first time, we use the adapted Hungarian Method to find an initial solution.We show that proposed method is superior to state-of-art existing algorithms.We suggest some directions for fruitful future research. This paper presents a penalty-based construction algorithm for the no-wait flow shop scheduling problem with the objective of minimizing makespan and total flow time of jobs. The proposed method, derived from Hungarian penalty method originally used for the classic assignment problem is employed to generate an initial schedule of jobs, which is further improved by an insertion technique to obtain an optimal or near-optimal schedule. The results of computational experiments on a large number of test problems show that the proposed method performs significantly better than the state-of-the-art procedures while requiring comparable computational effort.

An Improved Cuckoo Search Algorithm for Parallel Machine Scheduling

This paper proposes a cuckoo search-based algorithm (CSA) to minimize makespan for identical parallel machine scheduling problems. Job permutation schedules are used to implement CSA for this problem. We present a heuristic approach based on mod function to convert a continuous position in CSA into discrete permutation schedule for obtaining a cuckoo by Levy flights. Empirical results with a large number of randomly generated benchmark problem instances demonstrate that the proposed method produces solutions that are fairly superior to that of two state-of-the-art algorithms in the literature.

Machine-part cell formation for maximum grouping efficacy based on genetic algorithm

Cellular manufacturing system (CMS) makes use of application of group technology. The objective of cell formation problems (CFP) in CMS is to identify part families and machine cells in order to minimize the intercellular movement and to maximize the machine utilization within a cell. Previous study in CFP generally focused on maximizing grouping efficacy (GC) by minimizing exceptional elements as well as void elements. In this paper, a genetic algorithm heuristic is presented. Computational experiments were carried out with 20 benchmark problem sets. Computational results show that the proposed heuristic has shown to produce solutions in terms of GC that are either better than or competitive with the existing algorithms.

Comparison of Heuristics for Identical Parallel Machine Scheduling

This paper addresses the problem of scheduling n independent jobs processed nonpreemtively on m identical parallel machines with the objective of minimizing makespan. Since these scheduling problems are well known to be NP-hard, among various solution methodologies, heuristics are preferred most. They guarantee near-optimal solutions and due to their polynomial time algorithms require reasonable computational effort, especially for solving large problem sizes. We consider three popular heuristics, multifit, combine and listfit since the seminal work of McNaughton in 1959. We present different experimental frameworks to investigate these heuristics for a comprehensive comparative performance evaluation. We show, through computational experimentation, that listfit outperforms the multifit and combine heuristics in most of the problem instances, however, at the cost of increased time complexity. The computational results also reveal that the combine heuristic performs better than the multifit heuristic, while requiring almost similar computational effort.