We propose a new approach to the problem of searching a space of policies for a Markov decision process (MDP) or a partially observable Markov decision process (POMDP), given a model. Our approach is based on the following observation: Any (PO)MDP can be transformed into an "equivalent" POMDP in which all state transitions (given the current state and action) are deterministic. This reduces the general problem of policy search to one in which we need only consider POMDPs with deterministic transitions. We give a natural way of estimating the value of all policies in these transformed POMDPs. Policy search is then simply performed by searching for a policy with high estimated value. We also establish conditions under which our value estimates will be good, recovering theoretical results similar to those of Kearns, Mansour and Ng [7], but with "sample complexity" bounds that have only a polynomial rather than exponential dependence on the horizon time. Our method appl...

Gradient-based approaches to direct policy search in reinforcement learning have received much recent attention as a means to solve problems of partial observability and to avoid some of the problems associated with policy degradation in value-function methods. In this paper we introduce � � , a simulation-based algorithm for generating a biased estimate of the gradient of the average reward in Partially Observable Markov Decision Processes ( � s) controlled by parameterized stochastic policies. A similar algorithm was proposed by Kimura, Yamamura, and Kobayashi (1995). The algorithm’s chief advantages are that it requires storage of only twice the number of policy parameters, uses one free parameter � � (which has a natural interpretation in terms of bias-variance trade-off), and requires no knowledge of the underlying state. We prove convergence of � � , and show how the correct choice of the parameter is related to the mixing time of the controlled �. We briefly describe extensions of � � to controlled Markov chains, continuous state, observation and control spaces, multiple-agents, higher-order derivatives, and a version for training stochastic policies with internal states. In a companion paper (Baxter, Bartlett, & Weaver, 2001) we show how the gradient estimates generated by � � can be used in both a traditional stochastic gradient algorithm and a conjugate-gradient procedure to find local optima of the average reward. 1.

Cooperative games are those in which both agents share the same payoff structure. Valuebased reinforcement-learning algorithms, such as variants of Q-learning, have been applied to learning cooperative games, but they only apply when the game state is completely observable to both agents. Policy search methods are a reasonable alternative to value-based methods for partially observable environments. In this paper, we provide a gradient-based distributed policysearch method for cooperative games and compare the notion of local optimum to that of Nash equilibrium. We demonstrate the effectiveness of this method experimentally in a small, partially observable simulated soccer domain. 1 INTRODUCTION The interaction of decision makers who share an environment is traditionally studied in game theory and economics. The game theoretic formalism is very general, and analyzes the problem in terms of solution concepts such as Nash equilibrium [12], but usually works under the assu...

We describe a new approximation algorithm for solving partially observable MDPs. Our bounded policy iteration approach searches through the space of bounded-size, stochastic finite state controllers, combining several advantages of gradient ascent (efficiency, search through restricted controller space) and policy iteration (less vulnerability to local optima).

Solving partially observable Markov decision processes (POMDPs) is highly intractable in general, at least in part because the optimal policy may be infinitely large. In this paper, we explore the problem of finding the optimal policy from a restricted set of policies, represented as finite state automata of a given size. This problem is also intractable, but we show that the complexity can be greatly reduced when the POMDP and/or policy are further constrained. We demonstrate good empirical results with a branch-andbound method for finding globally optimal deterministic policies, and a gradient-ascent method for finding locally optimal stochastic policies.

Standard value function approaches to finding policies for Partially Observable Markov Decision Processes (POMDPs) are generally considered to be intractable for large models. The intractability of these algorithms is to a large extent a consequence of computing an exact, optimal policy over the entire belief space. However, in real-world POMDP problems, computing the optimal policy for the full belief space is often unnecessary for good control even for problems with complicated policy classes. The beliefs experienced by the controller often lie near a structured, low-dimensional manifold embedded in the high-dimensional belief space. Finding a good approximation to the optimal value function for only this manifold can be much easier than computing the full value function. We introduce a new method for solving large-scale POMDPs by reducing the dimensionality of the belief space. We use Exponential family Principal Components Analysis (Collins, Dasgupta, & Schapire, 2002) to represent sparse, high-dimensional belief spaces using low-dimensional sets of learned features of the belief state. We then plan only in terms of the low-dimensional belief features. By planning in this low-dimensional space, we can find policies for POMDP models that are orders of magnitude larger than models that can be handled by conventional techniques. We demonstrate the use of this algorithm on a synthetic problem and on mobile robot navigation tasks. 1.

In order for an agent to perform well in partially observable domains, it is usually necessary for actions to depend on the history of observations. In this paper, we explore a stigmergic approach, in which the agent’s actions include the ability to set and clear bits in an external memory, and the external memory is included as part of the input to the agent. In this case, we need to learn a reactive policy in a highly non-Markovian domain. We explore two algorithms: sarsa(λ), which has had empirical success in partially observable domains, and vaps, a new algorithm due to Baird and Moore, with convergence guarantees in partially observable domains. We compare the performance of these two algorithms on benchmark problems.

Partially observable Markov decision processes (POMDPs) provide a natural and principled framework to model a wide range of sequential decision making problems under uncertainty. To date, the use of POMDPs in real-world problems has been limited by the poor scalability of existing solution algorithms, which can only solve problems with up to ten thousand states. In fact, the complexity of finding an optimal policy for a finite-horizon discrete POMDP is PSPACE-complete. In practice, two important sources of intractability plague most solution algorithms: large policy spaces and large state spaces. On the other hand,

Neuroevolution is currently the strongest method on the pole-balancing benchmark reinforcement learning tasks. Although earlier studies suggested that there was an advantage in evolving the network topology as well as connection weights, the leading neuroevolution systems evolve fixed networks. Whether evolving structure can improve performance is an open question. In this article, we introduce such a system, NeuroEvolution of Augmenting Topologies (NEAT). We show that when structure is evolved (1) with a principled method of crossover, (2) by protecting structural innovation, and (3) through incremental growth from minimal structure, learning is significantly faster and stronger than with the best fixed-topology methods. NEAT also shows that it is possible to evolve populations of increasingly large genomes, achieving highly complex solutions that would otherwise be difficult to optimize.

Existing algorithms for discrete partially observable Markov decision processes can at best solve problems of a few thousand states due to two important sources of intractability: the curse of dimensionality and the policy space complexity. This paper describes a new algorithm (VDCBPI) that mitigates both sources of intractability by combining the Value Directed Compression (VDC) technique [13] with Bounded Policy Iteration (BPI) [14]. The scalability of VDCBPI is demonstrated on synthetic network management problems with up to 33 million states.