0

I have met a very practical problem in robotics field. As I am EE background and not familiar with algorithms, I am seeking for help here.

There are n destinations, and the destinations are to be divided into two groups(group A and group B). There are also two robots, robot A and robot B. Each destination of group A must be visited by robot A at least once. Each destination of group B must be visited by robot B at least once. All the information is given, weights, directions and etc.

Questions:

How to calculate the division, s.t. the two robots travel the minimum distance summing up? How to calculated the division, s.t. the time that two robots finish visiting all the destinations is shortest?

richieqianle
  • 602
  • 2
  • 7
  • 20
  • You might try adapting techniques for the Traveling Salesman Problem to this setting (with 2 robots instead of 1): https://en.wikipedia.org/wiki/Travelling_salesman_problem. – D.W. Jun 30 '17 at 05:19

1 Answers1

0

My recommendation is to look at auction based techniques for task allocation for robots. The idea is that agents bid in a market to add destinations to their plan. Agents that can add new destinations to their plans for lower cost (shorter distance traveled) get awarded the task. Faster to compute solutions (using minimum spanning tree heuristic, for example instead of solving the more difficult TSP problem) tend to find higher cost solutions but will find reasonably good solutions for a large number of robots and a large number of destinations.

A good starting point is the reference: Dias, M. Bernardine, et al. "Market-based multirobot coordination: A survey and analysis." Proceedings of the IEEE 94.7 (2006): 1257-1270.

robot_chef
  • 86
  • 1
  • 5