​

Assignment 3: 1rst Presentation DD2438

Group 8

Saad Guessous & Peiheng Li

Current solution: Heuristic based navigation

• Use 2 scripts: defensive mode and offensive mode
• (Inspired from the initial tournament)
• Use an algorithm to dynamically switch from the defensive to the offensive mode when needed
• Eg. have more defensive agents if our team is winning, and vice versa
• Do all the heavy calculations in the beginning
• Calculating all the pair distances
• (O(N²) with N the number of nodes in the grid)

Offensive agent

Used heuristics:

• Distance to the closest food
• Distance to the closest capsule
• Current number of carried food by the agent (it’s not good to greed too much)
• Current score / Food left
• Enemy observations, if chased -> go back to the base

We compute a score using these values * weights for each of the 9 actions (up, left, up left, stop, …) and choose the action with the highest score

Defensive agent

Am I Scared?

Escape

Can I see enemy?

Chase it

Estimate Position

Yes

Yes

No

No

Heuristically Best Move

One-grid estimation

Defensive Criteria

Reached

Pure-pursuit to the desired escape position

Target the enemy

Balance between a couple of estimations

Short demo video

Next steps

• Implement defensive mode
• Implement the algorithm to dynamically switch from defensive to offensive
• Try ORCA to handle agents collisions
• Make the code run properly on client/server mode, it is currently working much better on host mode

Progress Status Week 1

• Comment to customer paying 250 000kr for the report:
• “All is well. We have spent a bit more time than planned, and progress is as expected”
• Planned Time spent: 25%
• (Out of the combined 200h)
• Actual Time spent: 10%
• Out of the combined 200h
• Actual Progress: 10%
• (estimate progress towards completing assignment)
• Risk of not completing assignment: 4%

Include this slide, but do not spend time discussing it at the presentation. Remember to also fill in the Progress Report Google Sheet