Adaptive Configuration of evolutionary algorithms for constrained optimization

In the literature, many different evolutionary algorithms (EAs) with different search operators have been reported for solving optimization problems. However, no single algorithm is consistently able to solve all types of problems. To overcome this problem, the recent trend is to use a mix of operators within a single algorithm. There are also cases where multiple methodologies, each with a single search operator, have been used under one approach. These approaches outperformed the single operator based single algorithm approaches. In this paper, we propose a new algorithm framework that uses multiple methodologies, where each methodology uses multiple search operators. We introduce it as the EA with Adaptive Configuration, where the first level is to decide the methodologies and the second level is to decide the search operators. In this approach, all operators and population sizes are updated adaptively. Although the framework may sound complex, one can gain significant benefits from it in solving optimization problems. The proposed framework has been tested by solving two sets of specialized benchmark problems. The results showed a competitive, if not better, performance when it was compared to the state-of-the-art algorithms. Moreover, the proposed algorithm significantly reduces the computational time in comparison to both single and multi-operator based algorithms.