Novel Adaptive Simulated Annealing Algorithm for Constrained Multi-Objective Optimization  

作　　者：Chuai Gang Zhao Dan Sun Li 

机构地区：[1]School of Information and Communication Engineering,Beijing University of Posts and Telecommunications,Beijing 100876,P.R.China [2]Key Laboratory of Universal Wireless Communication,Ministry of Education,Beijing University of Posts and Telecommunications,Beijing 100876,P.R.China [3]Beijing Key Laboratory of Network System Architecture and Convergence,Beijing University of Posts and Telecommunications,Beijing 100876,P.R.China

出　　处：《China Communications》2012年第9期68-78,共11页中国通信（英文版）

基　　金：supported by the Major National Science & Technology Specific Project of China under Grants No.2010ZX03002-007-02,No.2009ZX03002-002,No.2010ZX03002-002-03

摘　　要：In recent years, sinmlated annealing algo-rithms have been extensively developed and uti-lized to solve nmlti-objective optimization problems. In order to obtain better optimization perfonmnce, this paper proposes a Novel Adaptive Simulated Annealing （NASA） algorithm for constrained multi-objective optimization based on Archived Multi-objective Simulated Annealing （AMOSA）. For han-dling multi-objective, NASA makes improverrents in three aspects： sub-iteration search, sub-archive and adaptive search, which effectively strengthen the stability and efficiency of the algorithnm For handling constraints, NASA introduces corresponding solution acceptance criterion. Furtherrrore, NASA has also been applied to optimize TD-LTE network perform-ance by adjusting antenna paranleters; it can achieve better extension and convergence than AMOSA, NS-GAII and MOPSO. Analytical studies and simulations indicate that the proposed NASA algorithm can play an important role in improving multi-objective optimi-zation performance.In recent years,simulated annealing algorithms have been extensively developed and utilized to solve multi-objective optimization problems.In order to obtain better optimization performance,this paper proposes a Novel Adaptive Simulated Annealing (NASA) algorithm for constrained multi-objective optimization based on Archived Multi-objective Simulated Annealing (AMOSA).For handling multi-objective,NASA makes improvements in three aspects:sub-iteration search,sub-archive and adaptive search,which effectively strengthen the stability and efficiency of the algorithm.For handling constraints,NASA introduces corresponding solution acceptance criterion.Furthermore,NASA has also been applied to optimize TD-LTE network performance by adjusting antenna parameters;it can achieve better extension and convergence than AMOSA,NSGAII and MOPSO.Analytical studies and simulations indicate that the proposed NASA algorithm can play an important role in improving multi-objective optimization performance.

关 键 词：simulated annealing constrained rmlti-objective optimizaztion adaptive sub-iteration search-ing sub-archive PARETO-OPTIMAL 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] X703[自动化与计算机技术—控制科学与工程]