Compendium of the most representative algorithms in print---vector-valued dynamic programming, linear programming, policy iteration, the weighting factor approach---for solving
multi-objective Markov decision processes, with or without reward discount, over a finite or infinite horizon. Mifrani, A. (2024)