An incremental fuzzy controller for large dec-POMDPs

Developing Communication Strategy for Multi-Agent Systems with Incremental Fuzzy Model

“…In this paper, we extend our previous method presented in [12]. In that method, each agent makes use of an individual fuzzy rule-base to interact with the environment.…”

“…This cycle repeats over infinite steps. This type of MAS problems is properly modelled by infinite-horizon Dec-POMDP [12]. Fig.…”

“…Value function is another approach to represent the policy in infinite-horizon Dec-POMDP problems [14]. In our previous work [12] we have introduced an incremental method to learn a fuzzy model as a value function. It generates a compact fuzzy rule-base as a solution that offers scalability for large MAS problems.…”

“…Incremental methods are also employed to find a policy for infinite-horizon Dec-POMDP. An incremental reinforcement learning algorithm is presented in [12] to create a compact fuzzy model as a solution of large MASs.…”

“…Therefore, presenting fuzzy model is desirable to solve a MAS in previous studies [11]. In [12] an incremental fuzzy controller has been introduced to find a solution of large MASs.…”