A New Adaptive Real-Coded Memetic Algorithm

A new adaptive real-coded memetic algorithm has been developed for continuous optimization problems. The proposed algorithm utilizes an adaptive variant of continuous ant colony system for local search. Here new adaptive strategies are utilized for online tuning of the number of local search steps and the width of the search interval over each dimension of the search space. A new crossover scheme is also developed and utilized. The new algorithm has been examined with CEC 2005 benchmarks and compared with three other state of the art works in the field. The comparisons have showed relatively better results.