Optimized edge appearance probability for cooperative localization based on tree-reweighted nonparametric belief propagation

Nonparametric belief propagation (NBP) is a well-known particle-based method for distributed inference in wireless networks. NBP has a large number of applications, including cooperative localization. However, in loopy networks NBP suffers from similar problems as standard BP, such as over-confident beliefs and possible nonconvergence. Tree-reweighted NBP (TRW-NBP) can mitigate these problems, but does not easily lead to a distributed implementation due to the non-local nature of the required so-called edge appearance probabilities. In this paper, we propose a variation of TRWNBP, suitable for cooperative localization in wireless networks. Our algorithm uses a fixed edge appearance probability for every edge, and can outperform standard NBP in dense wireless networks.

Diffusion Gradient Temporal Difference for Cooperative Reinforcement Learning with Linear Function Approximation

We introduce a diffusion-based algorithm in which multiple agents cooperate to predict a common and global statevalue function by sharing local estimates and local gradient information among neighbors. Our algorithm is a fully distributed implementation of the gradient temporal difference with linear function approximation, to make it applicable to multiagent settings. Simulations illustrate the benefit of cooperation in learning, as made possible by the proposed algorithm.

Cooperative off-policy prediction of markov decision processes in adaptive networks

We apply diffusion strategies to propose a cooperative reinforcement learning algorithm, in which agents in a network communicate with their neighbors to improve predictions about their environment. The algorithm is suitable to learn off-policy even in large state spaces. We provide a mean-square-error performance analysis under constant step-sizes. The gain of cooperation in the form of more stability and less bias and variance in the prediction error, is illustrated in the context of a classical model. We show that the improvement in performance is especially significant when the behavior policy of the agents is different from the target policy under evaluation.