zhangzhehong/vb-mapf

Vein-based Coalition for Multi-Agent Formation Task

NOTE

this is the coding for vein-based pattern formation in version3.0.

The key

1. the comparison on four different ven-based structures.

2. run 36 * 11 trials on four vein-based formation structures to record makespan and average distance.

3. the rectify algorithm

• full coalition
• understaffed coalition
  • only compensate from the previous coalition
  • compensate from the previous and next coalition
  • only compensate from the next coalition
• glut coalition
  • only send out to the previous coalition
  • send out to the previous and next coalition
  • only send out to the next coalition

The description

1. Preprocess.py:

   • input: a shape image
   • output: a set of dot locations, ROI (min_height,max_height,min_width,max_width)

2. RandomPoints.py:

   • input: the number of rows, the number of columns, the number of required dots
   • output: the set of randomly distributed locations

3. VeinCoalitions.py

   • input: the set of locations, row, col, the typeof vein-based structure
     • type=1,slope=0;type=2,slope=pi/2;type=3,slope=pi/4;type=4,slope=3pi/4
   • output: a vein-based group of locations

4. formation-VB1.py

   • type=1,slope=0,the vein-based formation

5. formation-VB2.py

   • type=2,slope=pi/2, the vein-based formation

6. formation-VB3.py

   • type=3,slope=pi/4,the vein-based formation

7. formation-VB4.py

   • type=4,slope=3pi/4, the vein-based formation

8. video

   • Show three demonstrations

video.mp4