Download presentation
Presentation is loading. Please wait.
Published byHelen Lewis Modified over 9 years ago
1
F INDING P LANS FOR R EARRANGING R OBOTS IN - LIKE E NVIRONMENTS Pavel Surynek Charles University in Prague Faculty of Mathematics and Physics The Czech Republic
2
P ATH P LANNING FOR M ULTIPLE R OBOTS Input: Graph G=(V,E) and a set of robots R={r 1,r 2,...,r μ }, where μ<|V| each robot is placed in a vertex (at most one robot in a vertex) a robot can move into an unoccupied vertex through an edge (no other robot is allowed to enter the vertex) initial positions of robots... simple function S 0 : R →V goal positions of robots... simple function S + : R→V Task: Find a sequence of allowed moves for robots such that all the robots reach their goal positions starting from the given initial positions Pavel SurynekPlanSIG 2008
3
M OTIVATION FOR THE P ROBLEM Rearrangement of agents in tight space Automated control of heavy traffic Data transfer with limited size of the cache memory Pavel SurynekPlanSIG 2008
4
A S PECIAL CASE WITH - LIKE E NVIRONMENT ( - LIKE G RAPH ) G θ (a,b,c)=(V,E) μ=|V|-1 Relation to general graphs Decomposition to bi-connected components Cycle decomposition of a bi-connected component Last cycle with a handle represents a -like graph Pavel SurynekPlanSIG 2008 S0S0 S+S+ r1r1 r2r2 r3r3 r5r5 r4r4 r6r6 r7r7 r2r2 r3r3 r1r1 r5r5 r4r4 r6r6 r7r7 G (3,3,2) Example: Exchanging robots r 1 and r 3 (transposition of robots) a b c B1B1 B2B2 B3B3 B4B4 B5B5 B6B6 B6B6
5
T HEORETICAL F OUNDATIONS FOR A S OLVING M ETHOD Interpret arrangement of robots as a permutation Proposition 1 Any permutation over μ elements can be obtained as a composition of at most μ-1 transpositions. Proposition 2 Any even permutation over μ elements can be obtained as a composition of at most μ-1 rotations along a triple. Proposition 3 Rotation along a triple is always solvable in a -like graph; transposition is solvable, if the -like graph contains an odd cycle. Pavel SurynekPlanSIG 2008
6
A S OLVING M ETHOD – T HETA -BOX Pre-calculate off-line optimal solutions for transpositions and rotations along triples Compose a sub-optimal solution of the optimal solutions of special cases Use a fast alternative method if pre-calculated solutions are not available Produces solutions of higher quality ( shorter ) than existing methods (Kornhauser et at.,1984 – MIT method) Pavel SurynekPlanSIG 2008
7
C ONCLUSIONS AND F UTURE W ORK Special case of path planning for multiple robots in -like environments Compose a sub-optimal solution of the optimal solutions of sub-problems (Theta-BOX method) Produces shorter solutions than existing method Future: Improve a process of the search for optimal solutions of the special cases (transposition, triple rotation), currently – a variant of IDA* is used Pavel SurynekPlanSIG 2008
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.