Research On Hashing Technology Of Historical Calculation Data In Evolutionary Algorithm

Evolutionary algorithms has now been widely used in practical problems for its system modeling and spatial searching capability. However, there is a problem of repetitive computation for individual fitness in the process of evolutionary algorithms. Especially, when solving complex engineering problems, fitness calculations will expend lots of time. In order to solve the problem of repetitive computation for individual fitness, it is useful to save the data which has been calculated, and check historical data before calculating the fitness, then the data that has been calculated can be used directly. Through preserving historical calculation data can reduce real calculation times for individual fitness.In this paper, the key value of hash table is coded firstly aimed at historical data’s characteristics in the evolutionary algorithm, and the hash functions and conflict management is studied. An efficient access method for the historical calculation data is put forward. On this basis, the efficient use of historical calculation data is studied. Another effective method for expensive problem is fitness valuations strategies. For the commonly used queries neighborhood problem in the fitness valuations strategies of the evolutionary algorithm, a neighborhood query method based on hash bucket is proposed. Finally, he simulation experimental data and analysis is given base on the use of hash bucket for the different issues. The simulation experiments show that the hash mechanism of the efficient use of historical calculations data can effectively reduce the times of the fitness calculations and improve the efficiency of the algorithm.