CAJ | 학술논문

本文旨在找到一种改进的万有引力搜索算法，快速准确解决组合优化问题，以最小旅行距离为目标的经典商人旅行问题为例，使用实数编码，并将连续数值区间的位置变量映射到离散的组合排序，即可行的旅行方案；为了有效地避免算法陷入局部最优解，使用模拟退火算法进行局部扩展搜索，进一步提高解的质量；最后通过仿真对比，得出结合模拟退火算法的万有引力搜索算法具有收敛速度快，精度高的特点这一结论。
본문지재조도일충개진적만유인력수색산법，쾌속준학해결조합우화문제，이최소여행거리위목표적경전상인여행문제위례，사용실수편마，병장련속수치구간적위치변량영사도리산적조합배서，즉가행적여행방안；위료유효지피면산법함입국부최우해，사용모의퇴화산법진행국부확전수색，진일보제고해적질량；최후통과방진대비，득출결합모의퇴화산법적만유인력수색산법구유수렴속도쾌，정도고적특점저일결론。
This paper aims at working out an improved GSA to rapidly and accurately solve the problem of combination opti-mization with TSP as the example through using real coding. In order to avoid partial optimized solution, we use SA to implement partial expanding search, so as to further improve the quality of the solution. Finally, through simulated comparison, we come to the conclusion that the GAS combined with SA is a rapid and ac-curate algorithm.