Intelligent Parameter Identification for a High-Cycle Accumulation Model of Sand With Enhancement of Cuckoo Search Algorithm

This study presents a novel approach of intelligent parameter identification (IPI) for a high-cycle accumulation (HCA) model of sand, which reduces the subjective errors on manual parameter calibration and makes the use of the HCA model more accessible. The technique is based on optimization theory and adopts the cuckoo search algorithm (CSA). To improve search ability and convergence speed of CSA, several enhancements are implemented. First, the improved CSA (ICSA) incorporates quasi-opposition learning to expand the search space and replaces the original search strategy with a Cauchy random walk to enhance global search ability. Second, an adaptive scaling factor is introduced in the algorithm's control parameters to achieve a better balance between exploration speed and accuracy. Third, a dynamic inertia weight is used to balance the search between global and local spaces when generating new nest positions after abandoning old ones. The performance of the ICSA-based IPI approach is evaluated by comparing it with the original CSA-based IPI and manual calibration in determining the HCA model parameters. A comprehensive analysis is also conducted to assess the effectiveness and superiority of each improvement strategy introduced in the ICSA over the original CSA. All comparisons demonstrate that the proposed ICSA-based IPI method is more powerful and efficient in finding optimal parameters.