遗传算法在分布式O S2CFAR检测系统优化中的应用

针对一个3 传感器分布式O S2CFAR 检测系统, 本文分别使用了基本遗传算法和改进的遗传模拟退火算法、小生境遗传算法进行优化搜索, 给出了一组不同检测条件下的准最优搜索结果。结果表明, 对于这一优化问题, 遗传模拟退火算法和小生境遗传算法都具有较好的适应性, 其中小生境遗传算法在搜索质量、稳定性和搜索速度上相对更好一些, 是一种较理想的多传感器分布式O S2CFAR 检测系统参数优化算法。
关键词: 分布式检测;O S2CFAR; 遗传算法; 模拟退火算法; 小生境; 系统优化
Abstract　Fo r a dist ribu ted o rdered stat ist ics (O S) con stan t false alarm (CFAR) detect ion sys2 tem , the search ing of the op t im um detecto r param eters and detect ion perfo rm ance under com2 p licated detect condit ion s is a typ ical non linear op t im izat ion p rob lem. A 32stat ion dist ribu ted O S2CFAR detect ion system is op t im ized by u sing simp le genet ic algo rithm s, genet ic sim u lated annealing algo rithm s and n iche genet ic algo rithm s, respect ively. A set of quasi2op t im um detec2 t ion p robab ility resu lt s of the system under differen t detect condit ion s are given. Resu lt s show that the genet ic sim u lated annealing algo rithm s and n iche genet ic algo rithm s are all su itab le fo r th is op t im izat ion. Among the two algo rithm s, n iche genet ic algo rithm s have bet ter search quality, stab ility and are mo re effect ively, it is an ideal op t im um too l fo r the op t im izat ion of dist ribu ted O S2CFAR detect ion system s.
Key words: dist ribu ted detect ion; O S2CFAR; genet ic algo rithm s; sim u lated annealing algo2
rithm s; n iche; system op t im izat ion