Optimized Look-Ahead Tree Policies: A Bridge Between Look-Ahead Tree Policies and Direct Policy Search
This work addresses computational efficiency and performance in decision-making for AI and optimization domains, representing an incremental improvement by bridging two existing techniques.
The paper tackles the challenge of combining direct policy search (DPS) and look-ahead tree (LT) policies for sequential decision-making by proposing a hybrid scheme that uses DPS to learn a node scoring function, enabling smaller trees. It shows experimentally that this method outperforms pure DPS and LT policies on four benchmark domains, requiring fewer policy evaluations and being robust to perturbations.
Direct policy search (DPS) and look-ahead tree (LT) policies are two widely used classes of techniques to produce high performance policies for sequential decision-making problems. To make DPS approaches work well, one crucial issue is to select an appropriate space of parameterized policies with respect to the targeted problem. A fundamental issue in LT approaches is that, to take good decisions, such policies must develop very large look-ahead trees which may require excessive online computational resources. In this paper, we propose a new hybrid policy learning scheme that lies at the intersection of DPS and LT, in which the policy is an algorithm that develops a small look-ahead tree in a directed way, guided by a node scoring function that is learned through DPS. The LT-based representation is shown to be a versatile way of representing policies in a DPS scheme, while at the same time, DPS enables to significantly reduce the size of the look-ahead trees that are required to take high-quality decisions. We experimentally compare our method with two other state-of-the-art DPS techniques and four common LT policies on four benchmark domains and show that it combines the advantages of the two techniques from which it originates. In particular, we show that our method: (1) produces overall better performing policies than both pure DPS and pure LT policies, (2) requires a substantially smaller number of policy evaluations than other DPS techniques, (3) is easy to tune and (4) results in policies that are quite robust with respect to perturbations of the initial conditions.