An Intelligent Acquisition Method For Cellular Automata Transition Rules Based On Grey Wolf Optimization

Cellular Automata(CA)is a dynamic model of discrete systems in spatial-temporal,which includes cells,states,neighborhoods,transition rules,and time.It is an effective method for describing,understanding,and simulating complex geographic systems which are especially used in urban extension modeling.Transition rules are the core of of CA and directly determine the transition process of the cells.Ant colony algorithm,neural network algorithm,analytic hierarchy process,decision tree,particle swarm algorithm,cuckoo algorithm,bee colony algorithm are main intelligent acquisition methods of CA conversion.Among them,bio-intelligence algorithms can get better CA transition rules which can intelligently obtain the nonlinear relationship between spatial variables and urban expansion,and express the CA transition rules through "If-Then" format,which can be more clear and easy to understand the urban expansion process.Intelligent algorithms demonstrates advantages in the mining of CA transition rules while there are also have deficiencies,such as not being able to ensure that the global optimal solution of the local domain can be achieved at the same time.Grey Wolf Optimization as a new type of intelligent optimization algorithm for biota,which has strong global exploration and local optimization capabilities.We use the city expansion modeling as an application case to explore a method based on Grey Wolf Optimization Algorithm to intelligently obtain the transition rules of cellular automata(GWO-CA).The mathematical model algorithm,framework and specific operator of GWO-CA algorithm are discussed and prototype system is designed and implemented.In this paper,the case of Nanjing expansion simulation is used,the application of dynamic simulation urban of CA is verified,the method of decision tree and Logistic regression are compared in the end.The main contributions are as follows:(1)The method of gray wolf optimization algorithm and cellular automata expression method were summarizedSo far,we describe the concept,mathematics,model framework and optimization process of the grey wolf optimization algorithm.After that,we classify and summary the existing cellular automata model,and validate the grey wolf optimization method and the cellular automata method which have bottom-up similarity and consistency in model implementation method.Lastly,constructing the CA transition rules expression method of the grey wolf optimization algorithm,and providing theoretical and technical support for further researches.(2)An intelligent acquisition method for cellular automata conversion rules based on Grey Wolf Optimization algorithm was proposedBased on the basic theories of grey wolf optimization algorithm and cellular automata,this paper constructed a grey wolf optimized cellular automata transition rules,defined the grey wolf optimization target vector and objective function for the cellular automata transition rules,and designed the gray wolf optimization key operator for mining rules of cellular automata transition.(3)CA transition rules mining tool based on grey wolf intelligence was achievedMATLAB language was used to designed and developed the grey wolf intelligent mining tool for the transition rules of cellular automata.Due to this tool is universal,it is not only limited to the intelligent acquisition of the rules of cellular automata transition in urban expansion,but also can be used in grey wolf intelligent acquisition of geographic cellular automata conversion rules,which can provided a new tool for intelligent acquisition of cellular automata transition rules.(4)GWO-CA model for integrated Grey Wolf Optimization algorithm and cellular automata was constructedIn this paper,city expansion model as an application case was used and the GWO-CA model integrated gray wolf optimization algorithm and cellular automata was built.Finally,the urban extension model of Nanjing was realized and also were compared with NULL model,Logistic-CA and decision tree method.The results prove the correctness and feasibility of the grey wolf smart acquisition method of CA transition rules.