A NEW APPROACH FOR THE SOLUTION OF SINGULAR OPTIMUM IN STRUCTURAL TOPOLOGY OPTIMIZATION

In order to overcome the difficulties caused by singular optima,inthe present paper,a new method for the solutions of structural topology optimizationproblems is proposed.The distinctive feature of this method is that instead of solv-ing the original optimization problem directly,we turn to seeking the solutions of asequence of approximated problems which are formulated by relaxing the constraintsof the original problem to some extent.The approximated problem can be solvedefficiently by employing the algorithms developed for sizing optimization problemsbecause its solution is not singular.It can also be proved that when the relaxationparameter is tending to zero,the solution of the approximated problem will convergeto the solution of the original problem uniformly.Numerical examples illustrate theeffectiveness and validity of the present approach.Results are also compared withthose obtained by traditional methods.