Kuo-Ching Ying

Particle swarm optimization for parameter determination and feature selection of support vector machines

Support vector machine (SVM) is a popular pattern classification method with many diverse applications. Kernel parameter setting in the SVM training procedure, along with the feature selection, significantly influences the classification accuracy. This study simultaneously determines the parameter values while discovering a subset of features, without reducing SVM classification accuracy. A particle swarm optimization (PSO) based approach for parameter determination and feature selection of the SVM, termed PSO+SVM, is developed. Several public datasets are employed to calculate the classification accuracy rate in order to evaluate the developed PSO+SVM approach. The developed approach was compared with grid search, which is a conventional method of searching parameter values, and other approaches. Experimental results demonstrate that the classification accuracy rates of the developed approach surpass those of grid search and many other approaches, and that the developed PSO+SVM approach has a similar result to GA+SVM. Therefore, the PSO+SVM approach is valuable for parameter determination and feature selection in an SVM.

An ant colony system for permutation flow-shop sequencing

Ant colony system (ACS) is a novel meta-heuristic inspired by the foraging behavior of real ant. This paper is the first to apply ACS for the n/m/P/Cmax problem, an NP-hard sequencing problem which is used to find a processing order of n different jobs to be processed on m machines in the same sequence with minimizing the makespan. To verify the developed ACS algorithm, computational experiments are concluded on the well-known benchmark problem set of Taillard. The ACS algorithm is compared with other mata-heuristics such as genetic algorithm, simulated annealing, and neighborhood search from the literature. Computational results demonstrate that ACS is a more effective mata-heuristic for the n/m/P/Cmax problem.

Multiprocessor task scheduling in multistage hybrid flow-shops: an ant colony system approach

The hybrid flow-shop scheduling problem (HFSP) has been of continuing interest for researchers and practitioners since its advent. This paper considers the multistage HFSP with multiprocessor tasks, a core topic for numerous industrial applications. A novel ant colony system (ACS) heuristic is proposed to solve the problem. To verify the developed heuristic, computational experiments are conducted on two well-known benchmark problem sets and the results are compared with genetic algorithm (GA) and tabu search (TS) from the relevant literature. Computational results demonstrate that the proposed ACS heuristic outperforms the existing GA and TS algorithms for the current problem. Since the proposed ACS heuristic is comprehensible and effective, this study successfully develops a near-optimal approach which will hopefully encourage practitioners to apply it to real-world problems.

Applying hybrid meta-heuristics for capacitated vehicle routing problem

The capacitated vehicle routing problem (CVRP) is one of the most important problems in the optimization of distribution networks. The objective of CVRP, known demands on the cost of originating and terminating at a delivery depot, is to determine the optimal set of routes for a set of vehicles to deliver customers. CVRP is known to be NP-hard problem, and then it is difficult to solve this problem directly when the problem size is large. In this paper, a hybrid algorithm of simulated annealing and tabu search is applied to solve CVRP. It takes the advantages of simulated annealing and tabu search for solving CVRP. Simulation results are reported on classical fourteen instances and twenty large-scale benchmark instances. From simulation results, the proposed algorithm finds eight best solutions of classical fourteen instances. Additionally, the solutions of the proposed algorithm have also admirable performance for twenty large-scale benchmark instances. It shows that the proposed algorithm is competitive with other existing algorithms for solving CVRP.

Dynamic parallel machine scheduling with sequence-dependent setup times using an iterated greedy heuristic

Topics related to parallel machine scheduling problems have been of continuing interest for researchers and practitioners. However, the dynamic parallel machine scheduling problem with sequence-dependent setup times still remains under-represented in the research literature. In this study, an iterated greedy heuristic for this problem is presented. Extensive computational experiments reveal that the proposed heuristic is highly effective as compared to state-of-the-art algorithms on the same benchmark problem data set.

Sequencing single-machine tardiness problems with sequence dependent setup times using an iterated greedy heuristic

The single-machine tardiness problem with sequence dependent setup times is a core topic for scheduling studies. Tardiness is actually a difficult criterion to deal with, even in a relatively simple manufacturing system, such as a single-machine is strongly NP-hard. Motivated by the computational complexity of this problem, a simple iterated greedy (IG) heuristic is proposed to solve it. To validate and verify the proposed IG heuristic, computational experiments were conducted on three benchmark problem sets that included weighted and un-weighted tardiness problems. The experiment results clearly indicate that the proposed IG heuristic is highly effective as compared to the state-of-the-art meta-heuristics on the same benchmark instances. In terms of both solution quality and computational expense, this study successfully develops an effective and efficient approach for single-machine total tardiness problems with sequence dependent setup times.

Minimising makespan in distributed permutation flowshops using a modified iterated greedy algorithm

The distributed permutation flowshop scheduling problem (DPFSP) is a newly proposed topic in the shop scheduling field, which has important application in globalised and multi-plant environments. This study presents a modified iterated greedy (MIG) algorithm for this problem to minimise the maximum completion time among all the factories. Compared with previous approaches, the proposed algorithm is simpler yet more effective, more efficient, and more robust in solving the DPFSP. To validate the performance of the proposed MIG algorithm, computational experiments and comparisons are conducted on an extended benchmark problem set of Taillard. Despite its simplicity, the computational results show that the proposed MIG algorithm outperforms all existing algorithms, and the best-known solutions for almost half of instances are updated. This study can be offered as a contribution to the growing body of work on both theoretically and practically useful approaches to the DPFSP.

An enhanced ant colony optimization (EACO) applied to capacitated vehicle routing problem

In this paper, an enhanced ant colony optimization (EACO) is proposed for capacitated vehicle routing problem. The capacitated vehicle routing problem is to service customers with known demands by a homogeneous fleet of fixed capacity vehicles starting from a depot. It plays a major role in the field of logistics and belongs to NP-hard problems. Therefore, it is difficult to solve the capacitated vehicle routing problem directly when solutions increase exponentially with the number of serviced customers.The framework of this paper is to develop an enhanced ant colony optimization for the capacitated vehicle routing problem. It takes the advantages of simulated annealing and ant colony optimization for solving the capacitated vehicle routing problem. In the proposed algorithm, simulated annealing provides a good initial solution for ant colony optimization. Furthermore, an information gain based ant colony optimization is used to ameliorate the search performance. Computational results show that the proposed algorithm is superior to original ant colony optimization and simulated annealing separately reported on fourteen small-scale instances and twenty large-scale instances.

Metaheuristics for scheduling a non-permutation flowline manufacturing cell with sequence dependent family setup times

The broad applications of cellular manufacturing make flowline manufacturing cell scheduling problems with sequence dependent family setup times a core topic in the field of scheduling. Due to computational complexity, almost all published studies focus on using permutation schedules to deal with this problem. To explore the potential effectiveness of treating this argument using non-permutation schedules, three prominent types of metaheuristics-a simulated annealing, a genetic algorithm and a tabu search-are proposed and empirically evaluated. The experimental results demonstrate that in general, the improvement made by non-permutation schedules over permutation schedules for the due-date-based performance criteria were significantly better than that for the completion-time-based criteria. The results of this study will provide practitioners a guideline as to when to adopt a non-permutation schedule, which may exhibit better performance with additional computational efforts.

A sequential exchange approach for minimizing earliness-tardiness penalties of single-machine scheduling with a common due date

Abstract This study focuses on a class of single-machine scheduling problems with a common due date where the objective is to minimize the total earliness–tardiness penalty for the jobs. A sequential exchange approach utilizing a job exchange procedure and three previously established properties in common due date scheduling was developed and tested with a set of benchmark problems. The developed approach generates results better than not only those of the existing dedicated heuristics but also in many cases those of meta-heuristic approaches. And the developed approach performs consistently well in various job settings with respect to the number of jobs, processing time and earliness–tardiness penalties for the jobs.