Monte-Carlo Swarm Policy Search
نویسندگان
چکیده
Finding optimal controllers of stochastic systems is a particularly challenging problem tackled by the optimal control and reinforcement learning communities. A classic paradigm for handling such problems is provided by Markov Decision Processes. However, the resulting underlying optimization problem is difficult to solve. In this paper, we explore the possible use of Particle Swarm Optimization to learn optimal controllers and show through some non-trivial experiments that it is a particularly promising lead.
منابع مشابه
Power System Reliability Evaluation using a State Space Classification Technique and Particle Swarm Optimisation Search Method
It is well-known that the reliability evaluation of composite power systems is computationally demanding. This work introduces a state space classification (SSC) technique that classifies a systems state space into failure, success, and unclassified subspaces without performing power flow analysis. The SSC technique was developed based on calculating the maximum capacity flow of the transmissio...
متن کاملFinding the Needle in the Haystack with Heuristically Guided Swarm Tree Search
In this paper we consider the search in large state spaces with high branching factors and an objective function to be maximized. Our method portfolio, which we refer to as heuristically guided swarm tree search, is randomized, as it consists of several Monte-Carlo runs, and guided, as it relies on fitness selection. We apply different search enhancement such as UCT, look-aheads, multiple runs,...
متن کاملThe 6th AISB Symposium on Computing and Philosophy: The Scandal of Computation - What is Computation?
platforms of computation Matthew Spencer, Etienne B. Roesch, Slawomir J. Nasuto, Thomas Tanay and J. Mark Bishop Stochastic Diffusion Search applied to Trees: a Swarm Intelligence heuristic performing Monte-Carlo Tree Search Thomas Tanay, J. Mark Bishop, Matthew C. Spencer, Etienne B. Roesch and Slawomir J. Nasuto Toward a Unified View of Computation in Neural Systems: A Reply to Shagrir and Pi...
متن کاملLocating and Characterizing the Stationary Points of the Extended Rosenbrock Function
Two variants of the extended Rosenbrock function are analyzed in order to find the stationary points. The first variant is shown to possess a single stationary point, the global minimum. The second variant has numerous stationary points for high dimensionality. A previously proposed method is shown to be numerically intractable, requiring arbitrary precision computation in many cases to enumera...
متن کاملOffline Monte Carlo Tree Search for Statistical Model Checking of Markov Decision Processes
To find the optimal policy for large Markov Decision Processes (MDPs), where state space explosion makes analytic methods infeasible, we turn to statistical methods. In this work, we apply Monte Carlo Tree Search to learning the optimal policy for a MDP with respect to a Probabilistic Bounded Linear Temporal Logic property. After we have the policy, we can proceed with statistical model checkin...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
دوره شماره
صفحات -
تاریخ انتشار 2012