Interconnects parasitic extraction using modified Particle Swarm Optimization

Three new Particle Swarm Optimization approaches are proposed. We used these approaches to solve a Curve fitting problem for Parasitic Extraction Macro-modeling application. In the first proposed approach, Wiggling PSO (WPSO); we enforce the particles to vibrate in their motion towards the best position -instead of straight motion- to enlarge the scanning area. The second approach, Incrementally Social PSO (ISPSO); is utilizing a variable weight for the social term (xg-x). This variability enables changing the social relationship between the particles from highly repulsive to highly attractive. Finally, we proposed a new Control inspired approach, PID-PSO, where we dealt with the PSO motion as a process that needs a controller to be optimized. It is quite common to use PSO to tune PID parameters but in this context we used PID to tune PSO motion. The performances of these three proposed approaches were measured on an extensive real data sets and used along with the understanding of the physical problem to offer various explanations of the theoretical aspects of the new algorithms.