Simultaneous Generation and Improvement: A Unified RL Paradigm for FJSP Optimization

---

overview

---

Running

---

Creating the scripts FSJP instance data by running ./dataprocess/CreatInstance.

def main():
    '''
    Generate Random instance by parameters
    '''
    batch_size = 50
    num_jobs = 20
    num_mas = 5
    opes_per_job_min = int(num_mas * 0.8)
    opes_per_job_max = int(num_mas * 1.2)
    case = CaseGenerator(num_jobs, num_mas, opes_per_job_min, opes_per_job_max,
                         flag_same_opes=False, flag_doc=True,
                         path='../dataset/2005test/')
    for each_case_index in range(batch_size):
        case.get_case(each_case_index)

Training model by running ./DuelingDQN_Train_main.py. And all the training parameters can be modified in file ./config.json. All the training data will be saved in ./runs.

Evaluating the model by running ./evaluate.py. Experiments Ganttchart visualization will be saved in ./render_result.

FJSP simulator

Users can implemente their customize algorithm by using our the FJSP environment. All the dependent files are included in directory ./env and ./utils. You can create a FSJP environment by running:

dataset = '1005'    #the dataset must be located in ./dataset
env = FJSPEnviroment(dir_path='./dataset/' + dataset)
env.reset()

more functions about simulator can find in ./env/Environment.py