Research And Application Of Coyote Optimization Algorithm In Structural Optimization Design

Structural Design Optimization can help to reduce unnecessary material consumption and construction cost,by means of numerical optimization techniques to design material-efficient or cost-effective structures.Over recent decades,it has been a research hotspot in the field of civil engineering.Due to a large number of calculations and analysis on structural parameters involved,intelligent optimization algorithms is an important method for structural design optimization.A large number of optimization algorithms have been used in solving design optimization of truss structures.However,with continuously changing conditions in structures and increasing complexity of structure types and component materials,a higher qualification is needed for algorithms of structural optimization.There are still great challenges for rate of convergence and robustness of algorithms of structural optimization.On the other hand,higher performance of intelligent optimization algorithms have been developing constantly in the field of Computer Science,which provides potential possibility for civil engineering applications.By ultilizing Coyotes Optimization Algorithm into the optimal design of truss structures,the study provides a reference for improving the current structural optimization algorithms.This paper refer to the basic theory of the Coyote Optimization Algorithm,improves the optimal design for different types of truss structures,and then applies the improved algorithms into the optimal design for truss structures with discrete variable and continuous variable respectively.The study covers the following aspects:(1)The Coyote Optimization Algorithm is a stochastic algorithm based on swarm intelligence and evolution mechanism.The problem-solving process simulates the social structure of the coyote population who follows the natural evolution mechanism,and whose diversity is increased by experiences exchange by the coyotes.This paper applies the COA to the sizing optimization of truss structures.Whether it is a sizing optimization with continuous or discrete variables,both optimization results confirm that this optimizer has good robustness and rate of convergence.(2)Because of rate of convergence is slowly and unstable under high-dimensional conditions,by applying fusion strategy of Harmony Search Algorithm(HSA)to improve the performance of the COA,the paper proposes an improved Coyote Optimization Algorithm enhanced with Harmony Search(COAHS)for single-objective sizing optimization problems of the truss structures with continuous and discrete variable respectively.In order to verify the effectiveness of the COAHS,it has been applied to numerical examples of various classic truss structures.The results prove that the COAHS can improve the overall optimizing capacity with higher rate of convergence and robustness,which can be effectively applied to single-objective sizing optimization problems of truss structures.(3)When applying the COA to the single-objective shape optimization of truss structures,the paper proposes a coupling method which contains an integer code and real number code so as to solve the challenge caused by two different types of design variables.The hybrid coding can solve the singular stiffness matrix of truss structure and solve the optimization problem effectively.(4)Based on COAHS and the concept of Pareto set,this paper proposes an algorithm named multi-objective COAHS(MCOAHS)for the multi-objective optimal design problems.In the iterative process,while combining with offspring and individuals from parent coyote,the novel algorithm filters the parent coyote for next iteration by the fast non-dominated sorting strategy,and then utilizes the dynamic distance to maintain the diversity of the solutions.As a result,Pareto optimal is evenly distributed,and the local optimum can be avoided.(5)Combining with the test function,the proposed MCOAHS is utilized to analyze the convergence of the results,and then employed for the classical multi-objective truss optimization problem.The results of quantitative simulation show that the proposed algorithm demonstrates good convergence and can be applied to the multi-objective optimal design of truss structures.(6)This paper summarizes the numerical results of COA,COAHS and MCOAHS for the optimal design of truss,and looks forward to the future research on how to improve the algorithm for more robustness results.