In this paper we present a hybrid system combining techniques from symbolic planning and reinforcement learning. Planning is used to automatically construct task hierarchies for hierarchical reinforcement learning based on abstract models of the behaviours’ purpose, and to perform intelligent termination improvement when an executing behaviour is no longer appropriate. Reinforcement learning is...

We design and implement a software controller that can distribute energy in a power grid. First, locally-weighted regression is performed on past data to learn and predict the energy demand. Then, reinforcement learning is applied to determine how to distribute power given some state. We demonstrate that our software, in conjunction with smart grids and network communication systems, can automa...

Reinforcement Learning has been used for a number of years in single agent environments. This article reports on our investigation of Reinforcement Learning techniques in a multi-agent and adversarial environment with continuous observable state information. We introduce a new framework, two-player hexagonal grid soccer, in which to evaluate algorithms. We then compare the performance of severa...

Network integration studies try to assess the impact of future developments, such as the increase of Renewable Energy Sources or the introduction of Smart Grid Technologies, on large-scale network areas. Goals can be to support strategic alignment in the regulatory framework or to adapt the network planning principles of Distribution System Operators. This study outlines an approach for the aut...

In railway power supply systems using AC frequencies lower than the public grids of 50/60 Hz, high voltage AC overhead transmission lines in railway grid frequency are used as one measure of strengthening the systems. An increased resistance to overhead high voltage AC transmission lines, may motivate cables for future railway power systems. With the frequency of 50/60 Hz, reactive power produc...

For improving the efficiency of resource use in dynamic network environment, the computational model base on multiagent is adopted, and an effective learning model based on the reinforcement learning and the multi-level organization learning is proposed in this paper. A series of formal definitions, such as the dynamic network grid (DNG), the computing agent, the cooperation computing team and ...