Optimum design of truss structures using discrete Lagrangian method

Abstract

This paper studies the minimum weight design of truss structures with discrete variables using the discrete Lagrangian method (DLM). The theory of the DLM is briefly presented first, and parameters that influence convergence speed and solution quality for this method are investigated and discussed. A revised searching algorithm is proposed for solving truss optimization problems subject to stress, buckling stress, and displacement constraints. The feasibility of the DLM is validated by three truss design examples. The results from comparative studies of the DLM against other discrete optimization algorithms for representative truss design problems are reported to show the solution quality of the DLM. It is shown that the DLM can often find better solutions for truss optimization problems than methods reported in previous literature by using appropriate weighting for the objective function and changing speeds for the Lagrange multipliers.