2022-06-14 meeting agenda

[MichalisPanayides]

What has been done:

• OHHH CANADA 🐻🦌🍁
• Reinforcement Learning: 📈📉
  • More flooded models ✔️
  • Flood model once stable ✔️
  • No upper boundary ✔️
  • Plot all servers rates using with different starting μ's ⌚
  • Get a 3rd plot that's the weighted mean ratio of the rate to the max rate ⌚
  • Utility function 8 🤔

To be discussed:

• [x] Annual review
• [x] Reinforcement Learning:
  • Increasing arrival rate (individual plots)
  • Flooding the model
  • Plots with no upper bound
  • Include fatigue in utility function?

To be done:

• Simulated annealing of RL algorithm 🔢
• Thesis 😭
• Other things I want to do:
  • Direct formula for waiting time 🕥🧍
  • Closed form formula for blocking time ➡🚫
  • Closed form formula for steady states probabilities ☢
    
    
    
    
    
    

           All hail the holy meeting agenda

( ﾉ ﾟｰﾟ)ﾉ( ﾉ ﾟｰﾟ)ﾉ( ﾉ ﾟｰﾟ)ﾉ( ﾉ ﾟｰﾟ)ﾉ     ＼(ﾟｰﾟ＼)＼(ﾟｰﾟ＼)＼(ﾟｰﾟ＼)＼(ﾟｰﾟ＼)

[MichalisPanayides]

To be done:

• Claim money:
  • [x] Send EXP5 form to Paul
  • [x] Send EXP5 form and Receipts to Chris
• Annual review:
  • [ ] 5 slides (3 slides on what's been done + 1 slide on thesis structure + 1 slide on finishing timeline)
• Reinforcement learning algorithm:
  • [x] Plot all servers rates using different starting μ's (4 plots; one per server)
    • => Example 7 - main_results.ipynb - util_7_with_state_probs
  • [x] Get a 3rd plot that's the weighted mean ratio of the rate to the max rate
    • Ratio = (rate at state / max possible rate)
    • i.e. Plot = ratio * pi_(u, v) over all iterations
    • => Example 8 - main_results.ipynb - util_7_with_state_probs
  • [ ] Flooded system scripts
    • [ ] Start with flooded system and then fix it
    • [ ] Try different parameters - 100 different sets of parameters (with different num of servers)
      • Script 1: Run chill model and then flood it
      • Script 2: Run flooded model
  • [ ] Think about utility function 8 🤔
    • Look at Yueyang Zhong's paper
    • Present the key aspects to Vince and Paul
    • Social sciences - inefficiencies from learned behaviours
  • [ ] Simulated annealing
• Thesis:
  • [ ] Write stuff
• Twente seminar talk:
  • [ ] Polish slides