How to Solve It: Modern Heuristics
How to Solve It: Modern Heuristics
A component-based specification approach for embedded systems using FDTs
SAVCBS '05 Proceedings of the 2005 conference on Specification and verification of component-based systems
Artificial Intelligence with Uncertainty
Artificial Intelligence with Uncertainty
A hybrid real-time component model for reconfigurable embedded systems
Proceedings of the 2008 ACM symposium on Applied computing
Component engineering for adaptive ad-hoc systems
Proceedings of the 2008 international workshop on Software engineering for adaptive and self-managing systems
Professional Microsoft Robotics Developer Studio
Professional Microsoft Robotics Developer Studio
Programming microsoft® robotics studio
Programming microsoft® robotics studio
Artificial Intelligence: A Modern Approach
Artificial Intelligence: A Modern Approach
Computer-assisted creativity: Emulation of cognitive processes on a multi-agent system
Expert Systems with Applications: An International Journal
Hi-index | 0.00 |
Autonomous agents are often restricted by the programs that make up their ‘intelligence' because they are installed on the same hardware as the agent. Since intelligence is software and therefore abstract, it is possible to separate the components which ‘create' the agent's intelligence from the agent itself. The disembodiment of intelligence allows agents to access components that may not be suited to their hardware for physical reasons, such as storage capacity or computational complexity. It has been long established that humans find solutions to problems by dividing a problem into a series or smaller sub-problems [1; 2]. Using web services, ‘intelligence components' can be created which perform a simple generic task on behalf of a client agent. These components may be used in different combinations in order to create customized solutions for particular problems. Intelligence components may be distributed across servers in different locations, allowing other agents to benefit from the differing implementations. In addition, software may be updated remotely by updating individual components. The model is aimed at creating a repository of useful functionality which may enable intelligent agents to focus on the process of intelligence rather than processing individual environmental states. The solution presented demonstrates that an agent may access distributed components in order to control behaviour, taking into consideration that components themselves have no concept of the environment in which the requesting agent exists. The problem of translating a ‘model' solution into the environment-specific solution is therefore left to the agent.