en
0.25
0.5
0.75
1.25
1.5
1.75
2
Lagrange Dual Decomposition for Finite Horizon Markov Decision Processes
Published on Nov 30, 20113073 Views
Solving finite-horizon Markov Decision Processes with stationary policies is a computationally difficult problem. Our dynamic dual decomposition approach uses Lagrange duality to decouple this hard pr
Related categories
Chapter list
Dual Decomposition of Finite Horizon Markov Decision Processes00:00
Outline00:12
Problem Framework00:29
Markov Decision Processes - 100:30
Markov Decision Processes - 200:38
Markov Decision Processes - 300:41
Markov Decision Processes - 400:44
Markov Decision Processes - 501:23
Problem Framework01:40
Influence Diagrams02:17
Dual Decomposition - 103:05
Dual Decomposition - 203:06
Dual Decomposition - 303:51
Dual Decomposition - 404:17
Dual Objective05:05
Slave Problem05:31
Master Problem06:05
Algorithm Summary06:45
Resource Allocation07:16
Resource Allocation - An Example08:38
Experiments09:44
Comparisons09:46
Chain Problem10:06
Mountain Car10:20
Puddle World10:41
Results10:59
Summary & Future Work - 112:25
Summary & Future Work - 212:26