Faiz Ahyaningsih
The quadratic assignment problem is a combinatorial problem of deciding the placement of facilities in specified locations in such a way as to minimize a nonconvex objective function expressed in terms of flow between facilities, and distance between location. Due to the non-convexity nature of the problem, therefore to get a 'good' starting point is necessary in order to obtain a better optimal solution. In this paper we propose a combined strategy (random point strategy to get initial starting point and then use forward exchange strategy and backward exchange strategy to get 'optimal' solution). As a computational experience we've solved the problem of Esc 16b, Esc 16c and Esc 16h from QAPLIB. Finally, we present a comparative study between Combined Strategy and Data -Guided Lexisearch Algorithm. The computational study shows the effectiveness of our proposed combined strategy. © 2017 Author(s).
Department of Mathematic, State University of Medan (UNIMED), Jln Willem Iskandar Psr V Medan Estate, Medan, 20220, Indonesia
