Cross Entropy Optimization Method And Its Application In Satellite Reliability Design Optimization

Efficient optimization methods play an important role in improving the performance of satellite reliability design optimization.Satellite design is an important part of the overall design of the satellite,and plays an important role in shortening the design cycle,reducing design redundancy,and improving the whole-satellite performance such as the mechanical properties,the thermal Properties and electromagnetic compatibility.Reliability Based Design optimization with high computational cost.In view of the above challenges,the cross-entropy optimization search algorithm based on sampling is deeply studied and applied to the reliability design optimization problem of a small satellite for a certain earth observation.The main work of the dissertation includes the following contributions:1.Aiming at the shortages of traditional cross entropy optimization methods in terms of convergence speed and computational efficiency,an improved cross entropy search algorithm is studied.Due to the characteristics of high-dimensional,multi-peak,discontinuous or semi-continuous,mixed variables and highly nonlinear of the satellite overall design problem.And in dealing with this kind of problem,the classic entropy algorithm is with the problem of slow convergence due to the lack of intelligence in the sample distribution parameter updating mechanism and the high computational cost caused by the large number of samples required for the Monte Carlo sampling.Regarding the issues above,the Elite Weight Strategy(EW)and the adaptive sample size strategy(AS)are studied to accelerate the convergence rate of the algorithm and greatly reduces the number of calls to the objective function of the optimization problem,thus effectively improving the optimization efficiency.The validity of the algorithm is verified by unrestricted multimodal test functions and constrained satellite engineering examples.2.Aiming at the shortcomings of the cross-entropy method which uses the penalty function to deal with the constraint conditions in computational efficiency,a cross entropy optimization algorithm based on Markov chain Monte Carlo(MCMC)is studied.It concerned the problem that if the penalty function method was used to deal with the constraints of the multi-constraint optimization model,the cross-entropy algorithm is with the problem of low sampling efficiency,and large calculation cost.And a new Cross Entropy-Markov Chain Monte Carlo(CE-MCMC)hybrid optimization algorithm is proposed without using penalty function.The new algorithm makes the sample points completely within the feasible region and distribute along the constraint boundary with large probability.It is with strong global search ability,high sampling efficiency and low computational cost.The validity is verified by two-dimensional test functions and the satellite engineering examples.3.Based on the improved cross-entropy algorithm in this dissertation,a research on the application of satellite reliability design optimization is carried out..It uses the Sequential Optimization and Reliability Assessment(SORA)as a framework to decompose the two-level nested problem of optimization-reliability analysis of the traditional solving engineering problem into two subproblems as deterministic optimization and reliability analysis.Among them,the deterministic optimization sub-problem uses the improved cross-entropy algorithm proposed in this paper as an optimizer,which greatly improves the optimization efficiency.Finally,aiming at the reliability design optimization of one small earth observation satellite,the validity of the algorithm is verified and the optimal solution is obtained under the reliability constraint.In summary,this dissertation studies on the design optimization method involved in the satellite overall design and proposes a cross-entropy optimization algorithm based on MCMC sampling,which has a strong global optimization ability and greatly improves the optimization efficiency compared with the traditional cross entropy algorithm,effectively solves the problem of complex multi-constraints satellite optimization problem.Based on which,the cross entropy reliability design optimization method based on sequential optimization and reliability analysis is further proposed and applied to the overall multidisciplinary reliability optimization design of the satellite,and the feasibility and effectiveness of the algorithm are verified.The method can be applied to the design optimization of the reliability of complex systems in other fields with great theoretical and practical significance.