- Solving MDPs with L.P.
  * try to solve it with a single LP
  * formalize search algorithm
  * numerical stability
  * implementation
  * evaluation of results
  * what action models does it work for?

- Value Iteration/ Policy Iteration with max-norm projection
  * bound projection error
  * implementation
  * non-expansive projection
  * what action models does it work for?

- CSI
  * define problem
  * apply to VI / PI / LP solutions of MDP

- POMDPs (VI/PI)
  * what to do after DP step
  * multi-agent case
  * implementation

- Multi/Distributed Agents
  * what do we want to achieve?
  * Markov Games?

- Application of Factored MDPs

- Continuous Variables

- Relational MDPs

- Reinforcement Learning

- Subsumption Architecture

- Hierarchical Planning