Fuzzy Desirability Evaluation Structure for Multi-response Inference System Optimization using the Genetic Algorithm

Hekmat, Siavash

doi:10.22034/jaisis.2022.370770.1051

Fuzzy Desirability Evaluation Structure for Multi-response Inference System Optimization using the Genetic Algorithm

Document Type : Original Article

Author

Siavash Hekmat

Faculty of Industrial and Mechanical Engineering, Qazvin Branch, Islamic Azad University, Qazvin, Iran

10.22034/jaisis.2022.370770.1051

Abstract

Fuzzy rule-based systems are among the best techniques of modeling and solving real-world complex problems which include plenty of inputs and outputs. Emulating the reasoning process of a human expert is the main characteristic of these systems. Rule-based systems are also appropriate for other purposes including consulting, diagnosis, learning, decision support, design, planning, or research. The present study intends to find optimized solutions for multi-response problems having multiple conflicting objectives using a fuzzy inference method. The fuzzy outputs of the considered type of problems are evaluated here applying a proposed desirability mapping structure tailored for fuzzy responses. Since ordinary desirability functions are not applicable for fuzzy output variables according to the different possible cases, a customized desirability evaluation structure and defuzzification technique is proposed in this regard. Additionally, the genetic algorithm is applied to search among the input values that optimize the whole responses simultaneously. Eventually, the application of the model is described in a numerical example.

Keywords

Multi-Response Optimization

Multi-Criteria Decision-Making

Fuzzy Inference System

Fuzzy Expert System

Genetic Algorithm

Subjects