A Cross-Entropy Optimization Algorithm for Continuous Function Based on Improved Sampling

In this paper, the Cross-Entropy (CE) optimization algorithm for continuous functions is studied. Aiming at the problem of computational efficiency caused by the large number of samples in Monte Carlo sampling, the Dynamic Weight strategy (DW) and the Adaptive sample Size strategy (AS) are proposed. The former can get the weighting coefficient of each sample by measuring the difference between the elite sample and the best sample, which can speed up the convergence rate of the algorithm. The latter can greatly reduce the number of calls to the objective functions by building relationship between the sample size and the standard deviation of elite samples, which can effectively improve the optimization efficiency. Finally, the validity of the proposed algorithm is verified by four unconstrained multimodal standard test cases and two constrained practical project examples. The results show that with similar capability to obtain the global optimal solution, the number of calls to the objective functions is reduced by 73.44%, 69.44%, 47.33%, 77.88%, 40.21% and 25.95% respectively compared with the traditional CE algorithm.