抄録
This paper presents an algorithm to solve efficiently multi-constraint problems in shape optimization problems. In the case of applying the Lagrange multiplier method to the multi-constraint problems, the Lagrange multipliers are decided by the Kuhn-Tucker conditions. The Kuhn-Tucker conditions require putting the Lagrange multiliers with respect to inactive inequality constraints zero. Instead of putting them zero, this paper proposed an algorithm of multiplying sufficiently large value to the diagonal term of shape gradient matrix. Applying this algorithm to volume minimization problems with mean compliance constraints with respect to multi-loads, well convergent results were obtained.
