Research And Simulation Of Population’s Co-evolution In Ecology System

In ecology system, biotic population is a complex system because of thediversity of individual populations, the diversity of populations and the interaction ofpopulations. The theory of complex adaptive system is one of the main results ofresearch in science of complexity. Its main idea is that how adaptation buildscomplexity. In the complexity theory the adaptive agent is the key point. By learningand accumulating experience, the agent interacts with environment and other agents,and then changes the environment and other agents. The environment and state ofother gents will affect state of the agent. So the process forms a feedback ring. Andthe circle is the power to keep the complex adaptive system moving.The research of population’s co-evolution in ecology system is essentially theresearch in complex adaptive system. So we can use the methods to the complexadaptive system on the modeling of biotic population system, such as autonomyoriented computing method. With it we can observe how biotic population co-evolute,and understand that the interaction of biotic population is the power which can keepthe system running. To the research of population’s co-evolution in ecology system,most mathematicians and ecologists use the dynamic equation method. With thedynamic equation, they can research and understand the rule in the population’sevolution by finding the key to the dynamic equation. With this method, we can’tobviously observe the rule in the population’s evolutionary process. But we use abottom-up approach around the external behavior and internal behavior of individualentities and observe the emergent phenomenon. With the bottom-up approach, it’s anew exploration in the population’s co-evolution process.The interaction of entities is a gambling process. So this paper introduces thelearning theory in games. In the beliefs-based learning model, fictitious play is awidely used learning model. This paper proposes a novel neighborhood correlatedempirical weighted algorithm which adopts players’ own strategies and theirneighborhood information, which is a neighborhood correlated empirical weightedalgorithm for fictitious play(NCEWFP). The comparison experiment results demonstrate that the neighborhood correlated empirical weighted algorithm canachieve a better convergence value. If using the NCEWFP to research the entities ofpopulation, it’s not much suitable. Aiming at the learning of entities of population,based on NCEWFP algorithm this paper proposes a new algorithm: CO-NCEWFP.With the CO-NCEWFP algorithm, this paper using the AOC idea to model bioticpopulation, proposes a biotic population framework. In the biotic populationframework, there are environment, population and resource. The key to theframework is how to model the population. In the individual learning mechanism,with the CO-NCEWFP algorithm, the entity has a better adaptive ability. At last thispaper introduces the biotic population to the predator-prey system. With the twogroup contrast experiments, it describes effectively the rules in population’sco-evolution and demonstrates that the framework works well.