Duration: 1:01:48

4,015 views

35

Many real-world problems require making decisions that involve multiple possibly conflicting objectives. To succeed in such tasks, intelligent agents need algorithms that can efficiently find different ways of balancing the trade-offs that such objectives present. In this talk, I provide an introduction to decision-theoretic planning in the presence of multiple objectives. First, I present an overview of multi-objective decision-theoretic formalisms and show that different assumptions about these problems lead to different solution concepts such as the convex hull and the Pareto front. Then, I present an algorithm developed in my group, called Optimistic Linear Support, which can be combined with any single-objective solver to yield an anytime algorithm that efficiently approximates the convex hull with bounded error.