File size: 1,961 Bytes
64d50f3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
---
library_name: hivex
original_train_name: WildfireResourceManagement_difficulty_2_task_0_run_id_2_train
tags:
- hivex
- hivex-wildfire-resource-management
- reinforcement-learning
- multi-agent-reinforcement-learning
model-index:
- name: hivex-WRM-PPO-baseline-task-0-difficulty-2
  results:
  - task:
      type: main-task
      name: main_task
      task-id: 0
      difficulty-id: 2
    dataset:
      name: hivex-wildfire-resource-management
      type: hivex-wildfire-resource-management
    metrics:
    - type: cumulative_reward
      value: 115.81460418701172 +/- 26.078782597878128
      name: Cumulative Reward
      verified: true
    - type: collective_performance
      value: 51.81407127380371 +/- 18.776135090696116
      name: Collective Performance
      verified: true
    - type: individual_performance
      value: 27.154735565185547 +/- 10.446210720821945
      name: Individual Performance
      verified: true
    - type: reward_for_moving_resources_to_neighbours
      value: 60.09906158447266 +/- 33.42797696299573
      name: Reward for Moving Resources to Neighbours
      verified: true
    - type: reward_for_moving_resources_to_self
      value: 2.2813201546669006 +/- 1.1529928776928962
      name: Reward for Moving Resources to Self
      verified: true
---

This model serves as the baseline for the **Wildfire Resource Management** environment, trained and tested on task <code>0</code> with difficulty <code>2</code> using the Proximal Policy Optimization (PPO) algorithm.<br><br>

Environment: **Wildfire Resource Management**<br>
Task: <code>0</code><br>
Difficulty: <code>2</code><br>
Algorithm: <code>PPO</code><br>
Episode Length: <code>500</code><br>
Training <code>max_steps</code>: <code>450000</code><br>
Testing <code>max_steps</code>: <code>45000</code><br><br>

Train & Test [Scripts](https://github.com/hivex-research/hivex)<br>
Download the [Environment](https://github.com/hivex-research/hivex-environments)