The ripple algorithm and its application in optimizing the parameters of the Muskingum model

This paper introduces a new heuristic intelligent optimization algorithm called the ripple algorithm (RA), inspired by the ripple phenomenon in nature. The algorithm mimics the structure of ripples and utilizes a three-layer ripple random point search starting from the central point. The search center group is determined based on the comparative information between the current fitness and the global optimal fitness. The expansion and contraction of ripples are controlled by the ripple radius function and contraction function, respectively, allowing the algorithm to converge towards the best point in the solution space independently. Through comparisons with other commonly used heuristic algorithms such as particle swarm optimization algorithm, standard genetic algorithm, and gravity search algorithm using test functions, the results demonstrate that the ripple algorithm exhibits strong competitiveness in lower dimensions. Moreover, when applied to parameter optimization of the Muskingum model, the ripple algorithm demonstrates excellent optimization ability, high accuracy, and practicality. The paper also discusses the ripple radius function, contraction function, and other parameters that impact the optimization process of the ripple algorithm, and proposes several ideas and directions for further improving the algorithm.