Hybrid Genetic Algorithm with Auto-tuning Parameters and K-mean Clustering Strategy for Multimodal Optimization

摘要

Evolutionary algorithms (EAs) have been successfully applied to solve many real-world optimization problems in the past decade, such as in manufacturing scheduling, transportation routing, maintenance scheduling or various design and operating optimization in the industries. This research proposes a hybridization algorithm based on genetic algorithm (GA) and differential evolution (DE) which are one of the EAs for solving a multimodal optimization problem. By combining GA and DE, it can significantly improves the overall performance of the algorithm on solving optimization problems. Fuzzy logic controller (FLC) is used to adaptively adjust crossover and mutation probabilities. We also incorporate a simple machine learning clustering method, called k-mean clustering method, by partition the solutions into clusters and extract the useful information from best cluster to assist in the search. We perform a benchmarking to test and demonstrate the efficiency of the proposed algorithm. The results show that the proposed hybrid GA/DE algorithm can archive better results comparing to the traditional EAs.