Shuffled Complex Evolution Model Calibrating Algorithm: Enhancing its Robustness and Efficiency

The Shuffled Complex Evolution (SCE-UA) has been used extensively and proved to be a robust and efficient global optimization method for the calibration of conceptual models. In this paper, we propose two enhancements to the SCE-UA algorithm, one to improve its exploration and another to improve its exploitation capability of the search space. A strategically located initial population is used to improve the exploration capability and a modification to the downhill simplex search method enhances its exploitation capability. This enhanced version of SCE-UA is tested, first on a suite of test functions and then on a conceptual rainfall-runoff model using synthetically generated runoff values. It is observed that the strategically located initial population drastically reduces the number of failures and the modified simplex search also leads to a significant reduction in the number of function evaluations to reach the global optimum, when compared to the original SCEUA. Thus, the two enhancements significantly improve the robustness and efficiency of the SCE-UA model calibrating algorithm.