Soft Computing, Part II
This is the second issue of a pair of two special issues on soft computing. The concept of soft computing was conceived by L. A. Zadeh in 1990. Soft computing refers to a coalition of methodologies, which are drawn together to deal with the pervasive imprecision in the real world. The principal members of this coalition include fuzzy logic, neuro-computing, evolutionary computing, probabilistic computing, chaotic computing, and machine learning. These coalition members are synergistic in the sense that, in general, better results can be obtained when they are used in combination rather than in stand-alone mode. Therefore, in the future most intelligent systems will be of hybrid type, employing a variety of soft-computing methodologies in order to achieve superior performance.
Within the past few years, soft computing began to grow rapidly in both theoretical developments and practical applications in various fields. This Special Issue aims to examine such growth in the field of operational management. The intended audience is primarily the practitioners of the industrial engineering profession who are responsible for managing the operations of a production system. This Special Issue is a companion to another issue dedicated to the technical aspects of realizing a product appeared in the same journal. For those who are more interested in the theoretical developments of soft computing methodologies, please refer to the journals and technical conferences that are devoted to these subjects.
This Special Issue is comprised of a total of eleven papers. The first seven papers address production planning and scheduling problems. The first paper is written by T. Chen. He formulated a new fuzzy linear programming (FLP) model that incorporates the philosophy of prioritizing demand classes for planning the mid-term production of a wafer production plant. For its solution, the FLP model is first converted into an equivalent LP model and then solved with an existing LP software. The new model is shown to achieve a higher value of the discounted cash flows than two reference models. The second paper is written by Chiu, et al. They proposed a case-based reasoning approach for estimating the due date of a wafer production order. By employing nonlinear similarity functions and dynamic weighting mechanisms, the proposed approach was shown to give more accurate due date estimations than back-propagation neural networks and conventional rules. The third paper is authored by H.F. Wang and K.-Y. Wu. They formulated a mixed integer-programming (MIP) model for the multi-period, multi-product, and multi-resource production scheduling problems. For its solution, a two-phase approach was developed. In phase 1, the search space of the MIP model is transformed into a preliminary pattern by a heuristic mining algorithm so that a hyper assignment problem can be formed as a reference model to be solved. In phase 2, a stochastic global optimization procedure that incorporates a genetic algorithm with neighborhood search techniques is designed to find the optimal solution.
The fourth paper is written by Yi and Wang. They proposed a fuzzy logic embedded genetic algorithm to solve the problem for scheduling batch jobs on identical parallel machines. The objective is to minimize the total earliness-tardiness penalties. The paper authored by Pérez, et al. studies the applicability and behavior of several niching genetic algorithms in solving job shop scheduling problems. They consider the behavior of a niching GA in terms of its efficacy, its diversity in convergence, and its exploration and exploitation properties. The results of this study establish some guidelines for selecting the appropriate niching method based on the need of a desirable behavior. The paper written by Chan et al. presented a real-time fuzzy expert system for scheduling parts at multiple decision points in a simulated flexible manufacturing system. A decision point arises whenever seizing of one available production resource is necessary. In order to deal with the newly emerged distributed scheduling problems, Jia et al. presented a modified genetic algorithm, which is also capable of solving traditional scheduling problems. The modified GA employs two steps to encode a distributed scheduling problem. The first step is to encode the factory information and the second step is to encode all the jobs’ operations and their processing sequences.
The eighth paper is authored by A. Turkcan and M. S. Akturk, in which a problem space genetic algorithm was used to solve bi-criteria tool management and scheduling problems simultaneously in a flexible manufacturing system. The quality of the Pareto-optimal set was evaluated by using the performance measures developed for multi-objective optimization problems. The ninth paper is written by Chiu et al. They proposed the integrated use of engineering process control and neural network to address the issue of identifying process disturbances. The proposed method was compared favorably with another neural network method and two commonly used SPC charts: Shewhart chart and cumulative sum (Cusum) chart.
The last two papers were contributed by M. Gen and his associates. A. Syarif and M. Gen developed a hybrid spanning tree-based genetic algorithm to solve exclusionary side-constrained transportation problems. In order to improve the performance of the genetic algorithm, a fuzzy logic controller was used to dynamically regulate the genetic operators. The proposed method was shown to be better than other conventional methods and the spanning tree-based genetic algorithm approach as a whole. Yun et al. wrote the last paper to investigate the performance of several hybrid genetic algorithms for three complex optimization problems in terms of three measures: the best fitness, the average fitness, and the CPU time.
The Guest Editors of this Special Issue are very thankful to all contributing authors. We hope that this Special Issue will become a forum for the dissemination of the current developments and also for initiating new partnerships among researchers and practitioners. To facilitate this goal, we thus created a dedicated webpage, which can be accessed at URL: http://www.csc.lsu.edu/trianta/Books/Special_Issue5/Special5.htm. Finally, the Guest Editors wish to express their gratitude to the Chief Editor of the Journal of Intelligent Manufacturing, Dr. Andrew Kusiak, for his support and guidance on this Special Issue.T. Warren Liao and Evangelos Triantaphyllou
Industrial and Manufacturing Systems Engineering Department (http://imse.lsu.edu)
Louisiana State University (http://www.lsu.edu)