Computing Optimal Policies for Attack Graphs with Action Failures and Costs.
Karel DurkotaViliam LisýPublished in: STAIRS (2014)
Keyphrases
- optimal policy
- average cost
- attack graphs
- markov decision processes
- initial state
- finite horizon
- attack graph
- state space
- decision problems
- reinforcement learning
- long run
- finite state
- multistage
- dynamic programming
- average reward
- sufficient conditions
- action space
- expected cost
- discounted reward
- infinite horizon
- average reward reinforcement learning
- serial inventory systems
- expected reward
- dynamic programming algorithms
- ordering cost
- partially observable markov decision processes
- inventory replenishment
- partially observable
- long run average cost
- data mining