Transient policies in discrete dynamic programming: Linear programming including suboptimality tests and additional constraints.
Arie HordijkLodewijk C. M. KallenbergPublished in: Math. Program. (1984)
Keyphrases
- dynamic programming
- linear programming
- optimal policy
- markov decision problems
- linear program
- partially observable markov decision processes
- lp relaxation
- optimal control
- state space
- steady state
- feasible solution
- lagrangian relaxation
- policy search
- integer programming
- markov decision processes
- column generation
- infinite horizon
- stereo matching
- primal dual
- discrete geometry
- discrete version
- objective function
- average cost
- optimal solution
- continuous state
- np hard
- test cases
- constraint propagation
- dynamic programming algorithms
- greedy algorithm
- statistical tests
- policy iteration
- markov decision process
- network flow
- reward function
- piecewise linear
- genetic algorithm
- reinforcement learning
- multiscale