A Fast, Network-based, Hybrid Heuristic for the Assignment of Students to Schools
针对学生分配学校问题，提出一种结合两种网络表示的快速混合启发式算法，解决传统线性规划模型因规模过大难以求解的难题，计算结果显示该方法非常高效。
In an era of declining and fluctuating enrolments, the determination of appropriate school sizes and student assignments poses difficult problems for administrators. For two decades, researchers have worked with variants of a linear programming districting model which minimizes total weighted distance as all students are assigned to schools constrained by both capacity and racial balance limitations. This model could aid in the generation of districting alternatives, but is difficult to solve in realistic applications with generalized solution procedures because of overall problem size. This has prompted the development of fast, network-based models, solvable with special algorithms and codes. We show, however, that these approaches all contain major shortcomings. We present a fast, hybrid heuristic based upon two network representations. Computational results demonstrate this to be a very fast solution approach for the general case of this districting problem.