Journal of Combinatorial Optimization ( IF 1 ) Pub Date : 2021-04-18 , DOI: 10.1007/s10878-021-00741-1 Adil Erzin , Georgii Melidi , Stepan Nazarenko , Roman Plotnikov
We consider a Two-Bar Charts Packing Problem (2-BCPP), in which it is necessary to pack two-bar charts (2-BCs) in a unit-height strip of minimum length. The problem is a generalization of the Bin Packing Problem. Earlier, we proposed an \(O(n^2)\)–time algorithm that constructs the packing of n arbitrary 2-BCs, whose length is at most \(2\cdot OPT+1\), where OPT is the minimum packing length. This paper proposes two new 3/2–approximate algorithms based on sequential matching. One has time complexity \(O(n^4)\) and is applicable when at least one bar of each 2-BC is greater than 1/2. Another has time complexity \(O(n^{3.5})\) and is applicable when, additionally, all BCs are non-increasing or non-decreasing. We prove the estimate’s tightness and conduct a simulation to compare the constructed packings with the optimal solutions or a lower bound of optimum.
中文翻译:
两张大条形图包装的3/2逼近
我们考虑两栏图打包问题(2-BCPP),其中有必要将两栏图(2-BCs)打包在最小长度的单位高度条中。问题是箱装箱问题的一般化。早些时候,我们提出了一种\(O(n ^ 2)\) -time算法,该算法构造n个任意2-BC的打包,其长度最大为\(2 \ cdot OPT + 1 \),其中OPT为最小包装长度。本文提出了两种新的基于顺序匹配的3/2近似算法。一个具有时间复杂度\(O(n ^ 4)\),并且在每个2-BC的至少一个小节大于1/2时适用。另一个具有时间复杂度\(O(n ^ {3.5})\)并且适用于所有BC不增加或不减少的情况。我们证明了估计的紧密性,并进行了仿真,以将构造的填料与最佳解决方案或最佳下限进行比较。