Research For The Improved AEA Algorithm Based On Cultural Algorithm And Its Application For The Model Parameter Estimation

AEA algorithm (Alopex-based evolutionary algorithm) is a kind of intelligent evolution algorithm based on the heuristic way of Alopex (algorithm of Pattern Extraction). This paper introduces Cultural algorithm(CA) into AEA algorithm,making use of CA double evolution mechanism to extract the implicit information of population space effectively. Then update the environment knowledge of reliability space constantly. Last, form a kind of mixed optimization algorithm CAEA that combined with CA structure and AEA evolution operation.Each iteration needs two generated populations to executive Alopex operation in AEA algorithms,so the quality of population of AEA algorithm has great effects on performance. In order to make full use of the population information, CAEA algorithm uses reliability space to extract and manage the first population elite and then generates the second population.In CAEA algorithm,environmental knowledge leads the entire population of the macro evolution each iteration.In population space,AEA algorithm search mode achieves individual micro evolution population.The population takes two evolutionary way both AEA evolution strategy and CA inspire way.So AEA algorithm can get elite individual,in addition to the influence of the individual independent variable for function value such as gradient search information. It's more advantageous to the evolution of species. AEA algorithm combined with CA double evolving structure has faster search speed,stronger convergence ability.This paper introduces the results of some typical test functions for testing AEA algorithm in the related paper,compares AEA algorithm with differential evolution algorithm, particle swarm algorithm and genetic algorithm according to function optimization results, convergence rate, convergence stability and so on.The results show that the performance of AEA algorithm is superior to the other algorithms.And then,compare CAEA algorithm with basic AEA algorithm through the 16 test function.The results show that the convergence speed, the accuracy of convergence of CAEA algorithm is greater than AEA algorithm.CAEA algorithm is used into heavy oil pyrolysis of model parameters optimization and growth curve model parameters optimization and compared with the data of the past papers. The simulation results show that CAEA algorithm can get higher precision and fitting smaller errors compared with the experiment data.