Instance dataset for a multiprocessor scheduling problem with multiple time windows and time lags: Similar instances...

The dataset presented in this paper introduces 384 new instances for the feasibility version of a multiprocessor scheduling problem with multiple time windows, positive time lags and exact time lags. The instances are constructed from subproblems in a logic-based Benders decomposition scheme introduced in "Logic-based Benders decomposition with a partial assignment acceleration technique for an avionics scheduling problem" (Karlsson, E., Rönnberg, E., Computers & Operations Research, 2022) [1]. A key aspect of the dataset is that even if two instances are highly similar, the computational performance of solving them with an IBM ILOG CP Optimizer model can be vastly different. There exists for example 44 pairs of instances with the same number of tasks and exact time lags, and the number of positive time lags differs with at most two, where one instance can be solved within 5 minutes and the other instance cannot be solved within 24 hours. Such differences make the instance dataset useful for investigating differences in computational performance of constraint programming solvers. The dataset can also be used to benchmark methods for multiprocessor scheduling. The dataset has been released under the Creative Commons Attribution 4.0 International license and can be used as it is or be adapted.