Abstract | ||
---|---|---|
Most exact algorithms for solving partially observable Markov decision processes (POMDPs) are based on a form of dynamic program- ming in which a piecewise-linear and convex representation of the value function is updated at every iteration to more accurately approximate the true value function. However, the process is computationally expen- sive, thus limiting the practical application of POMDPs in planning. To address this current limitation, we present a parallel distributed algo- rithm based on the Restricted Region method proposed by Cassandra, Littman and Zhang (1). We compare performance of the parallel algo- rithm against a serial implementation Restricted Region. |
Year | DOI | Venue |
---|---|---|
1999 | 10.1007/10720246_6 | ECP |
Keywords | Field | DocType |
pomdp solution,parallel algorithm,piecewise linear,value function | Dynamic programming,Mathematical optimization,Markov process,Computer science,Parallel algorithm,Partially observable Markov decision process,Algorithm,Markov decision process,Regular polygon,Bellman equation,Distributed algorithm | Conference |
Volume | ISSN | ISBN |
1809 | 0302-9743 | 3-540-67866-2 |
Citations | PageRank | References |
2 | 0.39 | 9 |
Authors | ||
2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Larry D. Pyeatt | 1 | 66 | 11.11 |
Adele E. Howe | 2 | 561 | 65.70 |