Fast simulated diffusion: an optimization algorithm for multi-minimum problems and its application to MOSFET model parameter extraction

A novel algorithm, fast simulated diffusion (FSD), is proposed to solve a multiminimal optimization problem on multidimensional continuous space. The algorithm performs a greedy search and a random search alternately and can find a global minimum with a practical success rate. An efficient hill-decending method for the greedy search is proposed. When the FSD is applied to a set of standard test functions, it shows an order-of-magnitude faster speed than the conventional simulated diffusion. An application of the FSD to a MOSFET parameter extraction problem is described.<
>