The Mimicry Of Physics Optimization Algorithm Based On Pd Control Research

Artificial Physics Optimization (APO) is a new algorithm for the rise of theintelligence, mainly used in function optimization, swarm robots searching filedand so on. The algorithm of APO simulates Newton law of motion, in which theindividuals attract or reject each other and change their speed following the totalforce from all other individuals, and go to a new place. The searching speed andsearching precision of the APO algorithm are all satisfying, but to some difficultoptimization problems, it also sometimes gets into the local best. To solve theproblem, in this thesis, the mathematical model of the APO algorithm isanalyzed; PD controller is introduced into the model of APO algorithm to adjustthe steps’ length fine and increase the ability of fine searching.The PD controller of the controlling system can reflect the trend of theincoming signal, and can forecast the change, which help the system keepingsteady. To help the algorithm to jump out of the local best, the PD controller isintroduced to the transfer function as a forward or backward part. And two newmodified algorithms are offered which names APO-PD1and APO-PD2. Theconvergence is proved in this thesis. The parameters Kp and Td are selectedthrough the uniform design method. Simulation results prove the efficiency ofAPO-PD1and APO-PD2through the comparison with PSO-PD1, PSO-PD2.APO, APO-PD1and APO-PD2are applied to the optimal coverageconfiguration of wireless sensor networks (WSN). The conclusion is that theperformance of APO-PD2is better than the other two algorithms on thecoverage rate of the wireless sensor network showed by the simulations.