| import os |
| import numpy as np |
| import pandas as pd |
|
|
|
|
| def table_plotting(model_names = ['PPO'],experiment_names = [''], execution_type = '', PI_proportion = True): |
| if PI_proportion: |
| df = pd.DataFrame(columns = ['Experiment','Model','All Costs (PI%)','Holding Costs (PI%)','Lost Sales (PI%)','Setup Costs (PI%)']) |
| else: |
| df = pd.DataFrame(columns = ['Experiment','Model','All Costs','Holding Costs','Lost Sales','Setup Costs']) |
| path = os.path.dirname(os.path.abspath('__file__')) |
| for experiment in experiment_names: |
| for model in model_names: |
| holding_costs_path = os.path.join(path,'results',model+'_'+experiment+'_'+'holding_costs'+'_'+execution_type+'.npy') |
| lost_sales_path = os.path.join(path,'results',model+'_'+experiment+'_'+'lost_sales'+'_'+execution_type+'.npy') |
| setup_costs_path = os.path.join(path,'results',model+'_'+experiment+'_'+'setup_costs'+'_'+execution_type+'.npy') |
| |
| all_holding_costs = np.load(holding_costs_path) |
| all_lost_sales= np.load(lost_sales_path) |
| all_setup_costs = np.load(setup_costs_path) |
| all_costs = all_holding_costs+all_lost_sales+all_setup_costs |
| if model == 'PI': |
| pi_all_holding_costs = all_holding_costs |
| pi_all_lost_sales_path = all_lost_sales |
| pi_all_setup_costs_path = all_setup_costs |
| pi_all_costs = all_costs |
| |
| if PI_proportion: |
| try: |
| df_temp = pd.DataFrame( |
| { |
| 'Experiment': experiment, |
| 'Model':model, |
| 'All Costs (PI%)': str(int(round(np.mean(all_costs - pi_all_costs)*100/np.mean(pi_all_costs),0)))+'$\pm$'+str(int(round(np.std((all_costs - pi_all_costs)*100/np.mean(pi_all_costs)),0))), |
| 'Holding Costs (PI%)': str(int(round(np.mean(all_holding_costs - pi_all_holding_costs)*100/np.mean(pi_all_holding_costs),0)))+'$\pm$'+str(int(round(np.std((all_holding_costs - pi_all_holding_costs)*100/np.mean(pi_all_holding_costs)),0))), |
| 'Lost Sales (PI%)': str(int(round(np.mean(all_lost_sales - pi_all_lost_sales_path)*100/np.mean(pi_all_lost_sales_path),0)))+'$\pm$'+str(int(round(np.std((all_lost_sales - pi_all_lost_sales_path)*100/np.mean(pi_all_lost_sales_path)),0))), |
| 'Setup Costs (PI%)': str(int(round(np.mean(all_setup_costs - pi_all_setup_costs_path)*100/np.mean(pi_all_setup_costs_path),0)))+'$\pm$'+str(int(round(np.std((all_setup_costs - pi_all_setup_costs_path)*100/np.mean(pi_all_setup_costs_path)),0))) |
| }, |
| index=[0] |
| ) |
| except: print('error') |
| else: |
| df_temp = pd.DataFrame( |
| { |
| 'Experiment': experiment, |
| 'Model':model, |
| 'All Costs': str(int(round(np.mean(np.sum(all_costs,0)),0)))+'$\pm$'+str(int(round(np.std(np.sum(all_costs,0)),0))), |
| 'Holding Costs': str(int(round(np.mean(np.sum(all_holding_costs,0)),0)))+'$\pm$'+str(int(round(np.std(np.sum(all_holding_costs,0)),0))), |
| 'Lost Sales': str(int(round(np.mean(np.sum(all_lost_sales,0)),0)))+'$\pm$'+str(int(round(np.std(np.sum(all_lost_sales,0)),0))), |
| 'Setup Costs': str(int(round(np.mean(np.sum(all_setup_costs,0)),0)))+'$\pm$'+str(int(round(np.std(np.sum(all_setup_costs,0)),0))) |
| }, |
| index=[0] |
| ) |
| df = df.append(df_temp) |
| |
| caption = 'Average cost and inventory level for each experimented model trained on 5 different seeds. One machine and two items scenario' |
| print( |
| df.to_latex( |
| index=False, |
| label='tab:coslidated_results', |
| caption=caption,escape = False |
| ) |
| ) |
|
|
