Utilizing the Unified Ant Colony Algorithm by Chaotic Maps

Ant colony optimization (ACOR) is a meta-heuristic algorithm for solving continuous optimization problems (MOPs). In the last decades, some improved versions of ACOR have been proposed. The UACOR is a unified version of ACOR that is designed for continuous domains. By adjusting some specified components of the UACOR, some new versions of ACOR can be deduced. By doing that, it becomes more practical for different types of MOPs. Based on the nature of meta-heuristic algorithms, the performance of meta-heuristic algorithms are depends on the exploitation and exploration, which are known as the two useful factors to generate solutions with different qualities. Since all the meta-heuristic algorithms with random parameters use the probability functions to generate the random numbers and as a result, there is no any control over the amount of diversity; hence in this paper, by using the best parameters of UACOR and making some other changes, we propose a new version of ACOR to increase the efficiency of UACOR. These changes include using chaotic sequences to generate various random sequences and also using a new local search to increase the quality of the solution. The proposed algorithm, the two standard versions of UACOR and the genetic algorithm are tested on the CEC05 benchmark functions, and then numerical results are reported. Furthermore, we apply these four algorithms to solve the utilization of complex multi-reservoir systems, the three-reservoir system of Karkheh dam, as a case study. The numerical results confirm the superiority of proposed algorithm over the three other algorithms.