一种新的蛋白质结构预测多模态优化算法  

Novel Multimodal Optimization Algorithm for Protein Structure Prediction

作　　者：程正华[1] 张贵军[1] 邓勇跃[1] 金媚媚[1] 

机构地区：[1]浙江工业大学信息工程学院,杭州310023

出　　处：《计算机科学》2013年第9期212-215,229,共5页Computer Science

基　　金：国家自然科学基金(61075062);浙江工业大学重中之重学科开放基金(20120811)资助

摘　　要：针对现阶段药物设计中对于蛋白质结构多模态的需求,提出了一种基于排挤差分进化策略的多模态优化算法。为了降低蛋白质构象空间求解的复杂度,算法采用能量极小化过程,有效缩小了可行域的搜索空间;同时,为了有效地平衡多模态优化问题的局部收敛性和模态多样性,在排挤差分进化算法的框架下,在保证算法收敛速度的前提下,算法采用空间局部性原理,同时随机选取不同交叉策略的集结思想又有效改善了种群的多样性。以脑啡肽为例,算法不仅得到了其全局最稳定结构,还获得了一系列局部最优结构。Aiming at the multimodal demand of protein structure for the drug design, a multimodal optimization algo-rithm based on differential evolution was proposed. In order to reduce the computation complexity of the protein confor-mational space, energy minimization.is applied to narrow the search space of feasible region. For balance local minima convergence and modal diversity of a multimodal optimization, under the framework of crowding differential evolution algorithm, under the premise of ensuring the convergence rate, the algorithm uses the principle of spatial locality and builds up procedures which randomly select a crossing strategy to increase the diversity of the population individual. Taking Met-enkephalin as benchmark, the new algorithm finds not only the global minimum energy conformation, but also many other distinct local minima.

关 键 词：差分进化算法 多模态优化 空间局部原理 集结过程 能量极小化 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]