Research On Reliability Optimization Design Of Structural System Based On Hybrid Teaching Learning Based Optimization

With the rapid development of economy,the safety of engineering structure has attracted more and more attention of scholars all over the world.Based on this situation,scholars put forward the idea of structural reliability based optimization design,which provides a new idea to solve structural design problems by using optimization algorithm.Because the optimization algorithm does not need an accurate model can also calculate better results.Teaching and learning optimization algorithm is a heuristic algorithm proposed in recent years,which has the advantages of few parameters and strong global search ability,but it also has some shortcomings and defects.In this paper,Teaching–Learning Based Optimization is improved in two different ways,and mixed with Tabu Search algorithm to improve the performance of teaching and learning algorithm.Finally,the reliability index is substituted into the optimization problem as a constraint condition,and the example shows that the hybrid algorithm can be applied to the structural reliability optimization problem.The main research contents and conclusions are as follows:(1)Firstly,Teaching–Learning Based Optimization is improved in two different ways,and then it is mixed with Tabu Search algorithm.Two hybrid algorithms,TLBOTS and CFTLBO-TS,are proposed.One is to improve the teaching factor in Teaching–Learning Based Optimization by nonlinear decreasing,the other is to introduce the contraction factor into the teaching stage of Teaching–Learning Based Optimization.Teaching–Learning Based Optimization based on the above two improved methods has the characteristics of fast global search speed and Tabu Search algorithm has strong local search ability.The test function shows that the two hybrid Teaching–Learning Based Optimization have the characteristics of fast convergence speed and better running results.(2)For the structural system reliability optimization problem,this paper proposes a structural probabilistic reliability optimization design model based on hybrid TLBO-TS algorithm.The model takes the structural mass minimization as a single objective function,and the probabilistic reliability index and cross-section product calculated by improved first-order second moment as constraints.The results of two examples of truss structure optimization show that the hybrid Teaching–Learning Based Optimization is better than the basic teaching and learning algorithm and the results of references.At the same time,the method has suitable for structural system reliability optimization.(3)For the structural non-probabilistic reliability optimization problem,this paper proposes a structural non-probabilistic reliability optimization design model based on CFTLBO-TS hybrid teaching and learning algorithm.The model is essentially a nested optimization design model,the outer layer is mass minimization,and the inner layer is non-probabilistic reliability analysis.The outer layer mass minimization is optimized by CFTLBO-TS hybrid algorithm,the inner layer non-probabilistic reliability analysis is optimized by iterative operation,and the non-probabilistic reliability index is calculated by interval model.The correctness and feasibility of the non-probabilistic reliability optimization design method based on CFTLBO-TS hybrid teaching and learning algorithm is verified by two examples of truss structure optimization...