Scherer Multidemand Multidimensional KP Instances

Test instance data for the 0-1 multidemand multidimensional knapsack problem (MDMKP) from the Primal Problem Instance Generator (PPIG) from Scherer et al. 2023. The instances allow for a greater range of correlation values between the profit coefficients and the coefficients for the knapsack and demand constraints. The test instances also flex the constraint tightness across each dimension. Providing a different structure than existing instances. The first 45 instances do not obey a predetermined correlation structure, while the remaining 135 instances are defined through an explicit predetermined correlation structure (Uniform(-1,1)) for each dimension of each coefficient used in the MDMKP.

The .txt file attached obeys the following structure, by line.

Problem number and best found objective value
number of variables, number of knapsack constraints, number of covering constraints
objective function coefficients
right hand side values knapsack constraints
right hand side values covering constraints
m rows of left hand side knapsack coefficients
n rows of left hand side covering coefficients

For any inquires contact the author at [e-mail]

本数据集为Scherer等人2023年提出的原始问题实例生成器（Primal Problem Instance Generator, PPIG）所生成的0-1多需求多维度背包问题（Multidemand Multidimensional Knapsack Problem, MDMKP）测试实例数据。该类测试实例可在收益系数与背包约束、需求约束系数之间实现更广范围的相关系数取值区间，同时可灵活调整各维度的约束紧密度，其结构与现有测试实例存在差异。其中前45个实例未遵循预设相关结构，剩余135个实例则针对MDMKP中各系数的每一维度，通过显式预设的Uniform(-1,1)相关结构进行定义。

附带的.txt文本文件按行遵循以下格式：
1. 问题编号与已找到的最优目标值
2. 变量数量、背包约束数量、覆盖约束数量
3. 目标函数系数
4. 背包约束的右端项取值
5. 覆盖约束的右端项取值
6. m行背包约束左端系数矩阵
7. n行覆盖约束左端系数矩阵

如有任何咨询，请联系作者：[e-mail]