(486k) Advanced Hybrid Sequential Niche Algorithm for Fining Multiple Solutions in Global Optimization

Detection of multiple solutions in optimization problems is important because multimodal objective functions are common in engineering and physics. For this reason, algorithms for finding all solutions in multimodal problems are desirable. Many kinds of global optimization which finds multiple optima have been developed; however, this is still challengeable work. W developed ?a hybrid niche algorithm', based on the sequential niche technique in combination with deterministic local optimizers (Moon & Linninger, 2009). In this method, we combined a genetic search with a deterministic method to accelerate locating multiple solutions in non-convex optimization problems. The method uses a self-adjusting niche radius to overcome the problem of clustering in the search space. Thus every niche area has a different size and updated. In this presentation, we will show an advanced version of our algorithm, which reuses the information obtained from prior candidate solutions to reduce search efforts. Also reasonable stopping criteria for this algorithm will be suggested in this presentation. To demonstrate our algorithm's performance and reliability, several challenging examples will be shown in our presentation.