An Interactive Descent Algorithm for Design Optimization of Composite Laminates Subjected to Multiple Loading Conditions

摘要

A model for designing strength controlled composite laminates and laminated structural elements is considered. The problem is formulated as a multicriterion optimization problem where laminate initial failure reserve factors with respect to the applied loading conditions are treated as criteria of the vector objective function. The original multiple criteria problem is reduced to a bicriterion problem by introducing parameters to combine linearly some criteria. An interactive descent algorithm is employed to solve the so-called Tchebycheff problem where the distance between a reference point (the global ideal criterion point) and the feasible objective region is minimized. The optimization procedure is defined through a point-to-set mapping, where at each cycle a new set of design points, i.e., laminate lay-up configurations and their mapping into the reduced criterion space are generated from one design point. The next state of the sequence is selected by the procedure and depends on the choice of the selected parameters. Convergence of the algorithm is considered with respect to the applied descent function. As each new point is generated by the algorithm the corresponding value of the descent function decreases in value, until the limit of the sequence is reached.