This project utilises the Multi-Start Genetic Algorithm Tabu Search (MS-GATS) method for the classic Flexible Job-Shop Scheduling Problem (FJSP), Extended-FJSP (EFJSP), and EFJSP with Transportation Constraints.
- "msts_algo_new.py" is the main entry point for this algorithm, and used for running the simulation.
- Certain parameters can be set in this file including:
- TS_cnt_max - Tabu Search Iteration Counter
- p_exp_con - Probability for choosing in e-greedy policy
- p_MA_OS - Probability of choosing MA or OS for critical path
- epochs - Maximum number of epochs
- eps_decay - Decay rate for e-greedy policy
- pop_size - Population size
- MA_algo_choice - Initial Dispatching Rules for machine assignment [Random, Greedy, LUM]
- OS_algo_choice - Initial Dispatching Rules for operation sequencing [Random, ERT, LRMT]
'data' directory has each dataset in individual directories. Including modified datasets with transportation times (T_times)
Virtual environment - use requirements.txt to create VEnv
Outputs are stored in 'output_models' named by date_PC-number.