用于函数优化的进化非选择算法实验分析

Experimental analyses of evolutionary negative selection algorithm for function optimization

进化非选择算法是将生物免疫系统的非选择机制和进化学习机制相结合而形成的算法，影响其求解效率的算子除了传统进化算法中的变异和选择算子外，还有非选择算子．通过函数优化实验验证了进化非选择算法的求解性能，结果表明非选择算子的引入使得进化非选择算法能够较好地跳出局部最优解，具有较为稳定的求解性能．与此同时，针对函数优化问题，给出了非选择算子相关的自我集大小和自我集每代更新数目这2个影响算法效率的重要参数的参考取值方法．

Evolutionary Negative Selection Algorithm is a synthesis of the negative selection mechanism and the evolutionary learning mechanism in biological immune system. In addition to mutation operator and selection operator of traditional Evolutionary Algorithm, negative selection operator will also affect the performance of the Evolutionary Negative Selection Algorithm. The function optimization experiments are conducted to demonstrate the performance of the Evolutionary Negative Selection Algorithm. And the experimental results show that with the negative selection operator, the Evolutionary Negative Selection Algorithm has better ability of escaping from local optimizations and getting a stable performance. At the same time, aiming at the function optimization problem, the empiristic methods of setting the self set size and the updating number of the self set at every generation are also given in this paper, both of which are important parameters about the negative selection operator, which will affect the performance of the algorithm very much.