Computational Intelligence Approach for Process Parameter Settings Using Knowledge Representation

Abstract

Abstract-This study proposes a fuzzy approach which integrates fuzzy rule sets in a chromosome. To enhance the functionality and capability of the fuzzy set, Genetic Algorithms (GA) technique is incorporated to produce a better and improved fuzzy set which is able to generate the expected result. Past data were selected to create the chromosomes and form the primary population set. This approach capitalizes on the merits of both techniques and offsets the drawbacks of them which may undermine the performance. This research signifies the hybrid approach to identify the optimal criteria for process control in order to achieve the target of the whole operations with an innovative methodology that has not been covered adequately to-date. A case example has been conducted to validate the practicality of the approach and the outcome demonstrated that the proposed approach is able to achieve the results as expected.Index Terms-Genetic algorithms, demand uncertainty, supply chain management, fuzzy logic model. I. INTRODUCTIONKnowledge and innovation are very important for potential business process improvements in a supply chain network with the increasing trends of short product life cycle and mass customization [1]. In order to improve the business process, lots of scholars focus on the process capability studies recently. The product features are measured and analyzed to determine the process ability process in order to meet product specifications [2]. Providing products and service in time with good quality is vital for business success. But how can a company optimize business processes to provide better products and service? Many scholars put great efforts in exploring the artificial intelligence (AI) techniques, such as intelligent systems [3].How to develop a next generation intelligent system emerges as a new research interest for AI research. For example, some scholars focus on the integration of Genetic Algorithms (GA) and Fuzzy Logic approach [4]. However, the application of this hybrid approach is limited. This study proposes a Fuzzy GA approach to integrate fuzzy rule sets and their associated membership function sets into GA approach. With this hybrid approach, both crisp and fuzzy data can be integrated into one chromosome which is impossible in traditional GA approach.The hybrid fuzzy-GA approach proposed in this study is conducted as follows. First, process parameter is encoded as a fuzzy rule sets and the fuzzy membership function is Manuscript received February 16, 2014; revised May 6, 2014. Henry C. W. Lau is with School of Business, University of Western Sydney, Australia (e-mail: H.lau@uws.edu.au).created for the fuzzy rule sets. Past data are selected to create the chromosomes and form the primary population set. Second, GA technique is used to produce a better and improved fuzzy set which is able to generate the expected result.This paper has 7 sections. In Section II, literature concerning the fuzzy logic approach, GA approach, and their related applications are presented. Section III is about t...