改进新解产生方式及温度函数的模拟退火算法.docVIP

改进新解产生方式及温度函数的模拟退火算法 　　摘要：简单介绍了传统模拟退火算法的流程、算法所涉及的重要参数、当下模拟退火算法改进的主要改进角度以及一种已有的改进算法――加温退火法。提出了一类基于改进新解产生方式及温度函数的模拟退火算法，一共包含四种新的改进算法，命名为：多粒子寻优模拟退火算法、混合温度模拟退火算法、混合多粒子寻优模拟退火算法、加温多粒子寻优模拟退火算法。最后分别将这四种改进算法应用于求解Sobol’g函数最小值和碎纸片拼接问题。实验证明改进后的算法是有效的，分别在解的质量以及算法效率上有所提升。关键词：模拟退火算法；新解产生方式；温度函数；Sobol’g函数；碎纸片拼接问题 　　中图分类号：TP18， TP30 文献标识码：A 文章编号：2095-2163（2015）05- 　　Simulated Annealing Algorithm based on Improving Production of New Solutions and Temperature Function 　　PENG Qiaozi， LU Yuting， LIN Yuyou， WANG Yingzhe 　　（School of Mathematics Science， Beijing Normal University， Beijing 100875， China ） 　　Abstract： The paper simply introduces the traditional simulated annealing algorithm through its process， key parameters， main aspects of improvement of the algorithm at present， and a new improvement named Simulated Annealing Algorithm with Heating Process which was put forward by other scholars. Then the paper puts forward a new type of simulated annealing algorithm based on improving production of new solutions and temperature function， including four improved algorithms which are named Multi-objectives Optimization Simulated Annealing Algorithm， Combined Temperature Simulated Annealing Algorithm， Combined Multi-objectives Optimization Simulated Annealing Algorithm and Multi-objectives Optimization Simulated Annealing Algorithm with Heating Process respectively. At last， the paper applies these four improved algorithms to determine the minimum value of Sobol’g function and restore the shredded paper respectively. The experiments demonstrate that the new type of simulated annealing algorithm is effective and show the improvement of both the solutions of those two problems and algorithm’s efficiency. 　　Key words： Simulated Annealing Algorithm； Production of New Solutions； Temperature Function； Sobol’g Function； Restoration of the Shredded Psaper 　　0 引 言 　　模拟退火算法 （Simulated Annealing Algorithm）是一种应用广泛的随机优化算法，最早的思想是由N. Metropolis等人于1953年提出。1983 年，S. Kirkpatrick等成功地将退火思想引入到组合优化领域