Problem Instances for the Generalized Assignment Problem (GAP) with Resource-Independent Task Profits and Identical Resource Capacity
Description
This dataset includes the problem instances generated for the Generalized Assignment Problem (GAP) with resource-independent task profits and identical resource capacity. We have set the four problem features - the number of tasks, the number of resources, the homogeneity of resources, and the relative capacity ratio of resources to total demands for tasks - with different levels for the features, resulting in 54 classes of the problem instances. For each class, 20 instances are generated, thus, a total of 1080 instances is included in the dataset. The main aim for this dataset creation was to test the performance of a solution algorithm to the target optimization problem and to characterize the performance of the algorithm as a function of the problem feature values. The findings along with the process could help a solution algorithm developer to understand the target optimization problem and thus design a solution algorithm with acceptable performance.
Files
Steps to reproduce
Please refer to "MATADATA.txt" inside of the zip file for the structure and contents of the dataset.