Data for: Solving the mixed model sequencing problem with reinforcement learning and metaheuristics
Description
The dataset contains the data used in our study "Solving the mixed model sequencing problem with reinforcement learning and metaheuristics". We provide three datasets: (1) The main dataset (folder "data") contains 2160 problem instances for the mixed model sequencing problem. The instances are based on the original dataset provided by Boysen (2011). Each instance contains information about processing times, cycle time, station length and demand plan. The dataset consists of ten mutations of the original dataset. Each mutation is a copy of the original dataset except for the demand plan. The demand plan is generated following a multinomial distribution. (2) The second dataset (folder "Robustness Check: Short-term disturbances") manipulates the demand plan of the main dataset. In each instance, the quantity of one model type is set to zero. The removed jobs are uniformly distributed over all other job models. (3) The third dataset (folder "Robustness Check: Added machines") adds machines to the initial instances. Each instance is extended by 1 to 5 machines. Original instances are used to provide the processing times for the new machines.