Differential Evolution Algorithm Using Population-Based Homeostasis Difference Vector

摘要

For the last two decades, the differential evolution is considered as one of the powerful nature inspired algorithm which is used to solve real-world problems. DE takes minimum number of function evaluations to reach close to global optimum solution. The performance is very good, but it suffers from the problem of stagnation when tested on multi-modal functions. In this paper, the population-based homeostasis difference vector strategy has been used to improve the performance of differential evolution algorithms. Here we propose two independent difference random vectors named as best difference vector and random difference vector which helps in avoiding stagnation problem of multi-modal functions. The performance of proposed algorithm is compared with other state-of-the-art algorithms on COCO (Comparing Continuous Optimizers) framework. The result verifies that our proposed population-based homeostasis difference vector strategy outperform most of the state-of-the-art DE variants.