Introducing Distance Factor Mimicry Physics Optimization Algorithm

APO， a global optimization algorithm inspired by the physics of mimicry,which as well is a kind of new intelligent optimal algorithm, is mainly used infunctional optimization, group robot system of distributed control and otherfields. The algorithm establishes rules of interaction between individuals toimplement optimization search and individual location update based onNewton’s second law. APO algorithm has been successfully applied to singleobjective search problem, besides, it has effected well in searching speed and theprecision. Yet when facing some particular difficult optimization problems, themethod will still be trapped in local optimization. Focusing on theabove-mentioned limitation, this article improves the rule of interaction inmimicry algorithm. On this condition, distance factor has been added on theintrinsic interaction discipline so that new interaction rule and individualdistance is related. As a result, the searching ability of algorithm is optimized.When the distance between individuals is larger than a certain value, theforce rule between individuals coincides with the rule of APO algorithminteraction. When the distance between individuals is smaller than a certainvalue, there exists only repulsion between individuals where the individual ofinferior adaptive value repulses the individual with superior adaptive value, nointeraction exists in the reverse direction. This model is put forward in order toavoid a local search algorithm which transforms the global search ability of thealgorithm. When the distance between individuals is larger than a certain value,the force between individuals follows the rule of APO algorithm interaction.When the distance between individuals is smaller than a certain value, thereexists only attraction between individuals where the individual of superioradaptive value attracts the individual with inferior adaptive value, no interactionexists in the reverse direction. This model is put forward in order to improve theprecision of the algorithm. Via several complex multimodal optimizationfunction simulation test, it is concluded that both the abovementioned modelsare effective. This article also has selected quality function between those twokinds of models and eventually gets power function as the quality function of these two kinds of models through the simulation test. In conclusion, the methodruns well.Applying APO, APO1and APO2to wireless sensor network coverageproblem and through numerical simulation experiment, it proves that wirelesssensor network coverage which APO2algorithm can achieve reaches mostextensively.