Adaptive Memory Tabu Search for Binary Quadratic Programs
提出自适应记忆禁忌搜索方法求解二元二次规划，实验表明该方法在求解难度最大的问题上能找到最优解，速度远超精确算法，且比现有最佳启发式方法更高效、解更优。
Recent studies have demonstrated the effectiveness of applying adaptive memory tabu search procedures to combinatorial optimization problems. In this paper we describe the development and use of such an approach to solve binary quadratic programs. Computational experience is reported, showing that the approach optimally solves the most difficult problems reported in the literature. For challenging problems of limited size, which are capable of being approached by exact procedures, we find optimal solutions considerably faster than the best reported exact method. Moreover, we demonstrate that our approach is significantly more efficient and yields better solutions than the best heuristic method reported to date. Finally, we give outcomes for larger problems that are considerably more challenging than any currently reported in the literature.