The Research Of Optimal Test Sequence Based On Improved Differential Algorithm

Recently,a number of important electronic devices such as rocket launching system,satellite missile systems,radar systems integration density become increased and its internal structure and functional complexity is also increasing.It puts forward higher technical requirements to the system fault detection and diagnosis.Test sequential design is the NP-complete.which must be solved in the process of fault diagnosis.Traditional test sequence optimization method has a lot of problems,such as the test time is too long,it is difficult to automatically generate the fault decision tree or unable to meet the test requirements,and it is not suitable for the fault diagnosis of complex systems.Differential evolution(DE)algorithm base its simple structure,less can tune parameters and is easy to actualize become very popular.In this paper,an improved differential evolution algorithm is used to design a test-fault matrix in a multi signal model system,obtained a Smaller fault diagnosis test cost and less test point set,has great significance to the practical engineering.The specific research work is as follows:(1)Firstly,estable an Mathematical model of the optimal test sequence base on Multi signal flow graph model,constructed the parameters of the test sequence problem five tparameter(S,P,T,C,D)model;Use difference algorithm to solve the problem,and improved the parameters of the algorithm.Without increasing the complexity of the algorithm,In order to find a balance between the global optimum and the convergence rate,propose a parameter adaptive dual mode differential evolution algorithm(IDDE).By constructing the dual mode mutation strategy,for each individual to increase the speed of inertia,test the benchmark function.The study shows that the improved algorithm can effectively reduce the sensitivity of the algorithm to the parameters and find the best ability,which provides a theoretical basis for the improvement of the differential algorithm.(2)Secondly,in order to improve the accuracy of the algorithm make it adapt to more and more complex problems constructed a fractal factor by dynamic correction de scaling factor F and cross factor CR new fractal algorithm differential evolution algorithm(FDE).Use benchmark function to test it and compare with currently recognized improved DE algorithm.The experimental results show that the proposed algorithm is effective and correct.By constructing a new fitness evaluation function,the optimal test sequence design for complex electronic system model is carried out.The experimental results show that the test point is reduced,and the testing cost is also reduced,Has practical engineering significance.(3)Finally,be aimed to the internal structure of large electronic equipment is too large and the fault isolation time is long and can't quickly generate the fault decision isolation tree.Propose a parallel hybrid algorithm based on the combination of DE algorithm and AO* algorithm DE-AO*.Using the characteristics of DE parallel optimization,we first select a set of test sets,then through the AO* powerful heuristic search capability and can automatically generate decision tree for each step of the test points to sort,The computational complexity of the system is reduced,and can effectively avoid the issue of "computational explosion".Applying the improved algorithm to the fault model of complex system,simulation results show that it can shorten the time of fault isolation and reduce the detection cost,which provides a feasible solution for the optimization of test sequences.