Title :
A VNS-ANT Algorithm to QAP
Author :
Han, Hongbo ; Ye, Jun ; Lv, Qiang
Author_Institution :
Soochow Univ., Taipei
Abstract :
Quadratic assignment problem (QAP) is one of the most difficult combinatorial optimization problems which has been tackled by many metaheuristic algorithms. This paper proposes a new algorithm, VNS-ANT, to solve QAP. The main idea is to hybridize a VNS (variable neighborhood search) mechanism and an ant algorithm (FANT). VNS-ANT enhances the local search in ANT with a VNS search in k-distance space guided by ant pheromone. In this way, VNS-ANT algorithm can not only efficiently speedup the improvement procedure in ANT, but sufficiently improve the veracity of VNS. The paper tests VNS-ANT on famous instances of QAPLIB. The experimental results show that VNS-ANT obtains better solutions than FANT with average 14.4% higher in solution quality.
Keywords :
combinatorial mathematics; quadratic programming; search problems; ant algorithm; ant pheromone; combinatorial optimization problems; metaheuristic algorithms; quadratic assignment problem; variable neighborhood search mechanism; Algorithm design and analysis; Computer science; Costs; Genetics; Simulated annealing; Statistics; Testing;
Conference_Titel :
Natural Computation, 2007. ICNC 2007. Third International Conference on
Conference_Location :
Haikou
Print_ISBN :
978-0-7695-2875-5
DOI :
10.1109/ICNC.2007.154
