Research On The Application Of The Improved Genetic Algorithm To The Multiple Rocket Attack Density Optimization

Multiple rocket is a kind of surface attack weapon, the shooting density of which is one of the important performance indexes. Through the in-depth study of launch dynamics, we find that it is an effective method to adjust the rocket firing sequence and shooting time is improve the multiple rocket fire intensity. According to the problem, an optimization model is built of which variable is the rocket firing time. And the problem is regarded as a multidimensional complex function optimization problem of no expression. What's more, the improved genetic algorithm is used to obtain the optimal solution.Function optimization is an important field of genetic algorithm, as well as an important means of algorithm performance testing. When optimizing the high-dimensional complex functions, the traditional genetic algorithm often premature or is in low efficiency of convergence. Based on the basic genetic algorithm, the study of the basic theory and the study of improving the real-coded genetic operators and operators portfolio strategy, this thesis puts forward a hybrid coding genetic algorithm (GA) and verifies the effectiveness of the new algorithm through the optimization of test function, which is applied in the multiple rocket firing density optimization to get better optimization results.Specific content as follows:(1) the study of the basic principle of genetic algorithm. This thesis is focused on the basic genetic algorithm, the algorithm of the basic elements of the design and evaluation.Besides, the real-coded genetic algorithm and sequential coding genetic algorithm are elaborated in detail in this paper.(2) Based on a certain type of multiple rocket fire simulation platform, placement of initial state parameters on the rocket was studied the multi-tube rocket is launching, including analyzing the correlation between the 22 parameters and rocket-propelled grenades in the initial state. Establish an optimization model with the rocket launch time as the variable(3) Put forward a selection operator with convergence and the ability of adaption.Through the study on the principle of selection operator and performance, put forward an adaptive selection operator updating in real time to make the individual only to produce offspring in the process of evolution between competition. Maintaining appropriate for each individual selection pressure is good for promoting the evolution of the population. Through the optimization of test functions and comparison with common selection operator, the effectiveness of the improved method can be verified.(4) The real-coded genetic algorithm in combination of crossover operator between several common effect were studied. Based on four kinds of crossover operator is used alone and serial, serial parallel combination of a total of 20 kinds of crossover operator performance are studied, provides reference for the combination and application of operator.(5) A hybrid genetic algorithm is put forward. this paper proposes a new coding method -series - real hybrid coding method, and designs the corresponding crossover mutation operator on the disadvantages of the limited ability of the real-coded global to make individuals be in good use in the process of evolution of population genetic information and to improve the efficiency of the algorithm. Through under different sizes of conditions, optimize the simple,complex, low dimension and high dimensional test function respectively, and compare them with the adoption of hybrid crossover - non-uniform mutation operator of the real genetic algorithm to verify the outstanding performance of the new algorithm in a small population size and in the high-dimensional complex function optimization.(6) Apply the proposed hybrid coding genetic algorithm to the optimal problem of multiple rocket firing density to get better effect. By analyzing the initial state parameters and the correlation analysis between rockets placement, establish the objective function with the initial disturbance as the variable. In the optimization process, in order to reduce the influence of random factors to determine the calculation objective function values, make random simulation calculation with the optimization results. It has been significantly improved compared with the original shot.