Year: 2016
Type: conference-paper
Abstract: Over the last two decades, different differential evolution (DE) variants have been successfully used to solve different optimization problems. However, no single DE algorithm has consistently been the best for solving a wide range of them. In the literature, this drawback has been tackled by using multiple DE operators in a single framework. However, utilizing a problem's landscape in the design of an efficient selection mechanism to emphasize the best-performing DE variant has not yet been thoroughly explored. Motivated by this fact, in this paper, a new two-phase (exploration and exploitation) multi-operator DE algorithm is proposed. It starts with the exploration phase, dynamically placing emphasis on the best-performing DE based on two landscape indicators and its performance history, and then repeats this process during the exploitation phase. To judge the performance of this algorithm, a variety of real-world optimization problems taken from different disciplines are solved. According to the results obtained, this algorithm shows superior performance to those of state-of-the-art algorithms.
Institutions: University of Canberra, UNSW Sydney, UNSW Canberra
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2588402173