【作者】 桑延超； 彭志红； 李谦；

【Author】 SANG Yan-chao,PENG Zhi-hong,LI Qian(School of Information Science and Technology,Beijing Institute of Technology,Beijing 100081,China)

【机构】 北京理工大学信息科学与技术学院；

【摘要】 提出一种粒子群算法和和遗传算法相结合的PSO-GA协同算法,将粒子群算法的全局最优引导能力引入到遗传算法的种群中来,使遗传算法能够合理利用PSO的快速搜索特性,从而提高遗传算法的收敛速度。同时利用遗传算法的全局收敛性,来避免粒子群算法容易陷入局部最优的缺点。通过共享全局最优的方法,两种算法得以实现协同操作。协同优化的同时,两种方法保有各自的独立性。因而两种方法可以以并行的方法执行,从而使优化速度进一步提高。该算法应用于低同源蛋白质三级结构预测,结果表明该算法优于单独的粒子群算法和遗传算法。相同的测试环境下,PSO-GA协同算法具有更快的收敛速度,优化结果也更接近实际的全局最优解。更多还原

【Abstract】 A new algorithm cooperating particle swarm optimization with genetic algorithm(PSO-GA)is presented to solve the problem of globe optimization.The best particle position pilot ability of PSO is introduced into this arithmetic to improve the convergence velocity of the solutions.At the same time,using the overall convergence character of GA,PSO can avoid running into local best position.By sharing the overall best solution,these algorithms cooperate.And at run time of algorithm,PSO and GA keep their own independence.So,they could be executed in parallel.And a better optimization velocity is obtained.In this study the PSO-GA is applied to low homology protein tertiary structure prediction.Through experiments in four true proteins,PSO-GA is proved to be better than PSO and GA in the same condition.It can achieve conformations which have lower energy values.And it does spend much less runtime than the other twos.更多还原