Okay so I have started working on three apporadches greedy , milp and rl agent based

I am more interested in seeing whether the rl based model works not 
For now the emphasis is given on the conflict resolution , scheduling proiority , and using platforms efficiently

Mean Reward: 41.94 Conflicts per Episode: 0.22 High Priority Success Rate: 97.0 % Avg Waiting Time: 1.84 Platform Utilization Balance: 0.255

This was the initial results. Here it can be seen that platform utilization is pretty less thats why lets penalize the model when it doesn't spread accross all the platforms and reward it upon using all platforms efficiently
Initially timestep was 50000, now i have made it 500000, which is around 245 iterations.

Now this is what I am getting 
Evaluation Results over 50 episodes Mean Reward: 49.0 Conflicts per Episode: 0.0 High Priority Success Rate: 100.0 % Avg Waiting Time: 0.0 Platform Utilization Balance: 0.25
Yuss yuss and this is clearly overfitting with 0 conflicts and 100 percent high priority success, not to mention the platform utilization is 0,25 that is it is using 1 out of 4 platforms regularly that is it is exploiting one platform for maximum returns 


Inorder to check whether it has overfitted or not , I created that file that will create random 50 timetable s and tested it and as I expected, this is what I get
--- Evaluating Timetable 1 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -161.0
Conflicts per Episode: 7.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 53.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 2 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -127.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 65.0
Platform Utilization Balance: 0.5
--- Evaluating Timetable 3 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -151.0
Conflicts per Episode: 6.0
High Priority Success Rate: 33.33333333333333 %
Avg Waiting Time: 56.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 4 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -146.0
Conflicts per Episode: 6.0
High Priority Success Rate: 33.33333333333333 %
Avg Waiting Time: 55.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 5 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -161.0
Conflicts per Episode: 6.0
High Priority Success Rate: 33.33333333333333 %
Avg Waiting Time: 72.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 6 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -106.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 45.0
Platform Utilization Balance: 0.5
--- Evaluating Timetable 7 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -181.0
Conflicts per Episode: 6.0
High Priority Success Rate: 20.0 %
Avg Waiting Time: 54.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 8 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -171.0
Conflicts per Episode: 7.0
High Priority Success Rate: 25.0 %
Avg Waiting Time: 64.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 9 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -106.0
Conflicts per Episode: 6.0
High Priority Success Rate: 100.0 %
Avg Waiting Time: 40.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 10 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -96.0
Conflicts per Episode: 5.0
High Priority Success Rate: 50.0 %
Avg Waiting Time: 55.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 11 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -71.0
Conflicts per Episode: 4.0
High Priority Success Rate: 60.0 %
Avg Waiting Time: 48.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 12 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -141.0
Conflicts per Episode: 6.0
High Priority Success Rate: 33.33333333333333 %
Avg Waiting Time: 66.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 13 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -151.0
Conflicts per Episode: 6.0
High Priority Success Rate: 50.0 %
Avg Waiting Time: 67.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 14 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -111.0
Conflicts per Episode: 5.0
High Priority Success Rate: 100.0 %
Avg Waiting Time: 43.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 15 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -146.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 51.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 16 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -146.0
Conflicts per Episode: 6.0
High Priority Success Rate: 50.0 %
Avg Waiting Time: 50.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 17 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -161.0
Conflicts per Episode: 6.0
High Priority Success Rate: 25.0 %
Avg Waiting Time: 66.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 18 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -161.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 66.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 19 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -111.0
Conflicts per Episode: 5.0
High Priority Success Rate: 100.0 %
Avg Waiting Time: 64.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 20 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -166.0
Conflicts per Episode: 7.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 71.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 21 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -146.0
Conflicts per Episode: 6.0
High Priority Success Rate: 25.0 %
Avg Waiting Time: 61.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 22 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -121.0
Conflicts per Episode: 5.0
High Priority Success Rate: 25.0 %
Avg Waiting Time: 52.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 23 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -151.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 69.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 24 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -16.0
Conflicts per Episode: 3.0
High Priority Success Rate: 50.0 %
Avg Waiting Time: 45.0
Platform Utilization Balance: 0.5
--- Evaluating Timetable 25 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -91.0
Conflicts per Episode: 4.0
High Priority Success Rate: 50.0 %
Avg Waiting Time: 42.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 26 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -141.0
Conflicts per Episode: 5.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 60.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 27 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -156.0
Conflicts per Episode: 6.0
High Priority Success Rate: 25.0 %
Avg Waiting Time: 65.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 28 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -166.0
Conflicts per Episode: 7.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 65.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 29 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -106.0
Conflicts per Episode: 6.0
High Priority Success Rate: 66.66666666666666 %
Avg Waiting Time: 44.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 30 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -146.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 54.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 31 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -146.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 60.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 32 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -146.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 68.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 33 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -136.0
Conflicts per Episode: 6.0
High Priority Success Rate: 33.33333333333333 %
Avg Waiting Time: 69.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 34 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -16.0
Conflicts per Episode: 3.0
High Priority Success Rate: 50.0 %
Avg Waiting Time: 39.0
Platform Utilization Balance: 0.5
--- Evaluating Timetable 35 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -191.0
Conflicts per Episode: 7.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 71.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 36 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -106.0
Conflicts per Episode: 5.0
High Priority Success Rate: 50.0 %
Avg Waiting Time: 52.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 37 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -156.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 73.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 38 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -166.0
Conflicts per Episode: 7.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 74.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 39 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -181.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 66.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 40 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -176.0
Conflicts per Episode: 7.0
High Priority Success Rate: 25.0 %
Avg Waiting Time: 64.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 41 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -131.0
Conflicts per Episode: 6.0
High Priority Success Rate: 33.33333333333333 %
Avg Waiting Time: 43.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 42 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -181.0
Conflicts per Episode: 7.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 68.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 43 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -186.0
Conflicts per Episode: 7.0
High Priority Success Rate: 25.0 %
Avg Waiting Time: 60.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 44 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -176.0
Conflicts per Episode: 7.0
High Priority Success Rate: 33.33333333333333 %
Avg Waiting Time: 87.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 45 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -131.0
Conflicts per Episode: 5.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 78.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 46 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -146.0
Conflicts per Episode: 6.0
High Priority Success Rate: 33.33333333333333 %
Avg Waiting Time: 56.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 47 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -96.0
Conflicts per Episode: 5.0
High Priority Success Rate: 100.0 %
Avg Waiting Time: 50.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 48 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -141.0
Conflicts per Episode: 6.0
High Priority Success Rate: 25.0 %
Avg Waiting Time: 68.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 49 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -161.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 72.0
Platform Utilization Balance: 0.25
--- Evaluating Timetable 50 ---
📊 Evaluation Results over 1 episodes
Mean Reward: -151.0
Conflicts per Episode: 6.0
High Priority Success Rate: 0.0 %
Avg Waiting Time: 69.0
Platform Utilization Balance: 0.25

Its definated overfitting 

Okay I think the main reason of this overfitting is that it has trained on a single timetable and got used to it so lets generalize it for 50 timetables first
Lets create a new file to generalize this 

Evaluation Results over 50 episodes
Mean Reward: 40.8
Conflicts per Episode: 0.0
High Priority Success Rate: 100.0 %
Avg Waiting Time: 0.0

Its again overfitting