Abstract:

A multiplier method based on activeset strategy was presented for solving optimization problems with many complicated constraints. We got a converted unconstrained problems, in which the maxvalue function could be approximated by the aggregate function. In each iteration procedure, only a small part of functions were usually involved, so the computation for gradient was significantly reduced. The numerical results show that the method is effective.

Key words: optimization with lots of constraints, multiplier method, activeset, aggregate function