The goal of the thesis is to increase the capabilities of existing risk-aware planning algorithm for Markov decision processes (MDPs) via Pareto front estimation. That is, the algorithm will compute estimates of the achievable risk-utility tradeoffs and use them to increase the quality of its decision making.

The expected contents of the thesis are the following:
- Extension of existing approaches to risk-aware MDPs with the capability to estimate risk-utility tradeoffs and to use these estimate to increase the quality of the algorithms' decision making.
- Implementation of the new algorithm in a stand-alone working tool, including an interface to work with user-defined MDP models.
- Thorough experimental evaluation of the new approach. This will include a comparison with the previous RAMCP/RAlph approached and with the Lagrangian-based method by Lee at al.
- The text part of the thesis will contain the thorough explanation of the new approach as well as the necessary explanation of the previous approaches on which the new ones build. Necessary technical preliminaries will be included. The thesis will discuss relevant related work from the area of risk-constrained planning and learning as well as a presentation and discussion of the experimental results.