Markov Decision Processes: Discrete Stochastic Dynamic Programming
Markov Decision Processes: Discrete Stochastic Dynamic Programming
Causal architecture, complexity and self-organization in time series and cellular automata
Causal architecture, complexity and self-organization in time series and cellular automata
Finding approximate pomdp solutions through belief compression
Finding approximate pomdp solutions through belief compression
Battery Model for Embedded Systems
VLSID '05 Proceedings of the 18th International Conference on VLSI Design held jointly with 4th International Conference on Embedded Systems Design
Solving POMDPs by searching the space of finite policies
UAI'99 Proceedings of the Fifteenth conference on Uncertainty in artificial intelligence
On the complexity of solving Markov decision problems
UAI'95 Proceedings of the Eleventh conference on Uncertainty in artificial intelligence
Multiscale Adaptive Agent-Based Management of Storage-Enabled Photovoltaic Facilities
Proceedings of the 2010 conference on ECAI 2010: 19th European Conference on Artificial Intelligence
The future of energy markets and the challenge of decentralized self-management
AP2PC'08 Proceedings of the 7th international conference on Agents and Peer-to-Peer Computing
| Hi-index | 0.00 |
We study the effects of different decision making schemes on theaccumulative rewards when photovoltaic (PV) facilities are intendedas a potential replacement for conventional peaking power plants.As the amount of solar irradiance usable by a PV module follows astochastic process, we compare the outcomes using the following twostrategies in a stochastic environment: (1) employing an optimaldecision making approach without any specific knowledge of theenvironment; and (2) optimal decision making based upon learningpatterns of the environment process. We examine the possibility ofintegrating a pattern learning approach --- called anε -Machine --- with a Partially Observable MarkovDecision Process (POMDP). This approach has been motivated in partby the fact that efforts in extending traditional learningapproaches to POMDPs have so far achieved only limited success. ThePV facility in our model consists of a PV panel and a battery, withan associated local, non-critical load. Under the assumption thatany PV generated power exceeding the maximum local consumptioncapacity must be dumped when the battery is full, the goal of theautonomous control agent is to maintain the maximum outputpotential to most effectively offset unexpected demand peaks, whileminimizing energy wastage in the presence of strong solarirradiance. The environment is assumed to follow a Markov processof a different order than the part of the system under theinfluence of the agent.