Consider the classical Bin Packing problem with $d$ different item sizes $s_i$ and amounts of items $a_i.$ The support of a Bin Packing solution is the number of differently filled bins. In this work, we show that the lower bound on the support of this problem is $2^{Ω(d)}$. Our lower bound matches the upper bound of $2^d$ given by Eisenbrand and Shmonin [Oper.Research Letters '06] up to a constant factor. This result has direct implications for the time complexity of several Bin Packing algorithms, such as Goemans and Rothvoss [SODA '14], Jansen and Klein [SODA '17] and Jansen and Solis-Oba [IPCO '10]. To achieve our main result, we develop a technique to aggregate equality constrained ILPs with many constraints into an equivalent ILP with one constraint. Our technique contrasts existing aggregation techniques as we manage to integrate upper bounds on variables into the resulting constraint. We believe this technique can be useful for solving general ILPs or the $d$-dimensional knapsack problem.
翻译:考虑具有 $d$ 种不同物品尺寸 $s_i$ 和对应数量 $a_i$ 的经典装箱问题。装箱问题解的支持度定义为不同填充方式的箱子数量。本文证明该问题支持度的下界为 $2^{Ω(d)}$,该下界与 Eisenbrand 和 Shmonin [Oper.Research Letters '06] 给出的 $2^d$ 上界在常数因子范围内匹配。此结论直接影响若干装箱算法的时间复杂度分析,包括 Goemans 和 Rothvoss [SODA '14]、Jansen 和 Klein [SODA '17] 以及 Jansen 和 Solis-Oba [IPCO '10] 提出的算法。为实现主要结论,我们开发了一种将多约束等式整数线性规划聚合为单约束等价整数线性规划的技术。该技术通过将变量上界整合至最终约束条件,与现有聚合方法形成鲜明对比。我们相信该技术对求解一般整数线性规划或 $d$ 维背包问题具有潜在应用价值。