1 / 13

The Freeze Tag Problem: How to wake up a swarm of robots

The Freeze Tag Problem: How to wake up a swarm of robots. B88505002 資訊三 曾宏偉 B88506043 資訊三 林明鴻 B88506065 資訊三 林宗茂. What ’ s is the freeze tag problem?. A set of n robots, modeled as points on vertices in some metric space. Initially, only one “ awake ” robot, and all others are asleep.

pabla
Download Presentation

The Freeze Tag Problem: How to wake up a swarm of robots

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. The Freeze Tag Problem: How to wake up a swarm of robots B88505002 資訊三 曾宏偉 B88506043 資訊三 林明鴻 B88506065 資訊三 林宗茂

  2. What’s is the freeze tag problem? • A set of n robots, modeled as points on vertices in some metric space. • Initially, only one “awake” robot, and all others are asleep. • Goal: to wake up all the robots a.s.a.p.minimize the “makespan”. • To wake a robot, an awake must go to its location; once awake, the new awaked robot can wake up other robots.

  3. Input: • A graph G=(V,E) l(ei) stands for the weight(length) of edge ei belongs to EThe cost of a robot passing ei; n(vi) stands for the number of robots on vertex vi belongs to V.

  4. Output: • Minimize to makespan of traversal all vertices using awakened robots.

  5. Facts • NP-hard, even for star graphs with an equal number of robots at each vertex.

  6. Greedy Algorithm for Star Graph • Scenario: Simplify the graph into star graph, with one awake robot a central node, and only one robot on each leaf. • Sort edges into l(e1)< l(e2)<……< l(en) • Go to the shortest edge first, and after visit a node, a new robot can awake other robots. • If all nodes are visited, done.

  7. P-time? Feasible Solution? • P-time? A: Of course! Since this is a greedy algorithm. • Feasible Solution? A: Sure, it really visits all node, and wake up all robots. The algorithm seems perfect, however…

  8. Consider the graph… 2k-1 edges with length 1 2k edges with length k 1 edge with length 3k

  9. What Greedy will do? • Go to the edge with length 1 first, and wake up all robots with distance to central node as 1, cost 2*lg2k=2k • Then the 2k to the edge with kcost k. • Finally go to the edge with length 3kcost k(for a r go back to central)+3k4k • Sum up to 7K. • However, this can be finish at 3k+4wake up a robot at edge length 1, at time 2, wake up the node with length 3k. At time 4, a robot go to edge with length k to wake robotssum up only to 3k+4.

  10. Wow! The ratio comes up to 7/3 • Is 7/3 the approximation ratio?

  11. Wow! The ratio comes up to 7/3(Cont.)

  12. A PTAS • Divide the edges into long and short sets L stands for the set of long edges. S stands for the set of short edges. T stands for (3/7)*Opt. Sol. of Greedy Alg.

  13. A PTAS (cont.) • 1. Completely Enumeration for L, to find the best wake up location for edges in L. • 2. Greedy for fill short edges into long edges. • result in an 1+ -approximation alg.

More Related