Global optimization of classified composite laminated structures based on genetic algorithms

A technique of global optimization aimed at composite laminated structures with high variety in distribution of stress was proposed. Under the consideration of connection and the efficiency of computation, the structure laminates were divided into regions with particular global layers and region layers on it. Special gene coding and operator were designed to match the design of division and genes of each region were treated separately in heredity. Combined with the finite element method (FEM), a subprogram written to simulate the stiff matrix of composite laminates based on classical lamination theory was used to analyze the model to get the deflection, weight and so on. as parameters of the object function and took the state whether the structure was failure or not judged by Tsai-Wu failure criterion as its condition. An example provided proves the optimized laminate structure reasonable and believable and having more efficiency in loading than one optimized in whole laminate.