Nature Inspired Approaches for Identification of Optimized Fuzzy Model: A Comparative Study

The identification of an optimized model is one of the key issues in the field of fuzzy system modeling. This has gained significant importance since; most of the real life systems are highly complex and nonlinear. Fuzzy model identification involves two stages i.e. identification of input and output membership functions as well as generation of rule base for the system being modeled. The fuzzy modeling or fuzzy model identification can be formulated as a search and optimization problem where the goal is to find the parameters of fuzzy model based on some evaluation criteria such that model gives optimal performance. Therefore, search and optimization techniques can be applied to the problem of model identification. Owing to their ability to deal with highly complex problems nature/biologically inspired approaches are currently amongst the most powerful algorithms suitable for fuzzy model identification problems. The research work reported in this paper is focused on five new approaches to model identification based on nature inspired optimization algorithms namely Biogeography Based Optimization Approach (BBO), Big Bang Big Crunch Optimization Based Approach (BB-BC), Artificial Bee Colony Optimization Based Approach (ABC), Ant Colony Optimization Based Approach (ACO) and Firefly Algorithm Based Approach (FA). These approaches have been implemented and validated successfully on two standard benchmark data sets to suggest robust, tractable and low cost solutions.