Anderson, J., and M. Evans, "A Generic Simulation System for Intelligent Agent Designs", Applied Artificial Intelligence, Volume 9, Number 5, October, 1995, pp. 527-562. -7-  Home/Search   Document Details and Download   Summary   Related Articles   Check

....and allow repeatability. To this end it should provide (i) libraries including at least one commonly used scheduling technique, like discrete time or event based simulation; ii) mechanisms to cluster agents in groups and apply different scheduling techniques to each group. According to [1], the environment of a problem can be represented by a collection of objects and agents. In our sense, a domain is defined by the environment and causal knowledge of how those objects and agents interact with one another. Domain facilities embrace two sub types of requirements: The first type ....

Anderson J. and Evans M. (1995). A Generic Simulation System for Intelligent Agent Designs. Applied Artificial Intelligent, v.9, n.5, pp. 527-562.

....to support controlled simulation worlds. A. Manage Scheduling Techniques: The platform should support controlled simulations and allow repeatability. It should provide libraries with the most usual scheduling techniques, like discrete time simulation and event based simulation. According to [1], the environment of a problem can be represented by a collection of objects and agents. In our sense, a domain is defined by the environment and causal knowledge of how those objects and agents interact with one another. Domain facilities embrace two sub types of requirements: The first type ....

Anderson J. and Evans M. A Generic Simulation System for Intelligent Agent Designs. Applied Artificial Intelligent, v. 9, n. 5, pp. 527-562, 1995.

....have used a single state variable to record the (real valued) position of each agent, obstacle and food item in the environment. This requires fewer state variables when the number of agents is small (and is the default approach adopted by several agent simulation systems, for example, Gensim [42] and SIM AGENT [41] but means that every state variable must be accessed in order to determine which objects in the environment (agents, obstacles or food items) are visible to an agent. Unless the state variables are grouped in some way based on their values (see, e.g. 5] 30] this means ....

John Anderson and Mark Evans, "A generic simulation system for intelligent agent designs," Applied Artificial Intelligence, vol. 9, no. 5, pp. 527--562, October 1995.

No context found.

Anderson, J., and M. Evans, "A Generic Simulation System for Intelligent Agent Designs", Applied Artificial Intelligence, Volume 9, Number 5, October, 1995, pp. 527-562. -7-

.... important elements of control for the evaluation of intelligent systems (Hanks et al. 1993) as well as solutions to problems in the real world that are beyond the capabilities of current AI technology, allowing research one area to proceed despite the immaturity of research in related areas (Anderson, 1995). Simulation also allows us to model natural intelligent systems, thereby deriving a better understanding of the techniques employed in those systems and ultimately the means to exploit those techniques in new applications (e.g. Picault and Collinot, 1998) Using a simulator to design and evaluate ....

.... to design and evaluate intelligent agents in realistic environments places enormous demands on a simulation tool: everything from supporting multiple agents and their interactions, to providing detailed control over trials in an environment, to accurate perception within computational bounds (Anderson, 1995). Surrounding these specific issues however, is the more pervasive problem of wide applicability: in order to perform ongoing research, where agent designs and the environments in which they are examined change as development pursues, we require a tool that will easily support such changes. ....

[Article contains additional citation context not shown here]

Anderson, J., and M. Evans, "A Generic Simulation System for Intelligent Agent Designs", Applied Artificial Intelligence, Volume 9, Number 5, October, 1995, pp. 527-562.

....is possible. Waiting may also affect the importance of the intention that brought about this activity (or intentions further back in abstraction) which may further alter these constraints. 4. Implementation and Summary We have implemented this architecture using the Gensim simulation testbed [3] under Macintosh Common LISP to provide a simulated world for the agent to inhabit. Gensim manages an object oriented representation of an environment (including the abilities of agents) presents object attribute descriptions of objects for agent perception (using a model of the agent s limited ....

Anderson, John, and Mark Evans, "A Generic Simulation System for Intelligent Agent Designs", Applied AI 9:5, 1995, pp. 527-562.

....anywhere where every possible contingency cannot be anticipated. This includes performing an activity with which we are not completely familiar, performing the activity in conjunction with others in the short or long term, or performing an activity in a different situation than is usual [Anderson, 1995]. In order to apply a routine effectively and flexibly in the face of greater variability than can be completely anticipated, we possess a vast collection of more general knowledge that allows us to integrate alternatives seamlessly with our routine. We can divert from our routine when it makes ....

....and resource dependencies for the purposes of clarity. However, constraints may be applied to a broad range of concepts, from physical restrictions and requirements [Fox, 1983] to expectations of actions or other agents and control of agent components such as memory retention and deliberation [Anderson, 1995; Evans et al. 1992] In addition to these, constraints within this architecture are used to represent direct preferences for resources, actions, or activities; restrictions on agent focus (to particular tasks, knowledge or particular perceptual information) to represent agent policies for ....

[Article contains additional citation context not shown here]

Anderson, John, and Mark Evans, "A Generic Simulation System for Intelligent Agent Designs", Applied AI 9:5, October, 1995.

....in this approach. 2. Improvisation The behaviour of a Waffler agent is based on the observed behaviour of humans in dealing with the extensive variability that occurs during the course of day to day activities through improvisation. Improvisation is observable in the bulk of human activities [Anderson, 1995; Hodgson and Richards, 1974; Jencks and Silver, 1972] and allows humans to respond flexibly and creatively by employing compiled routines supplemented with more general background knowledge in real time. During the course of an activity such as preparing a meal or driving to work, for example, ....

.... and relationships between entities (expectation, requirement, and temporal constraints) to abstract policies (behavioural goals and preference constraints) to representing normative responses to particular situations and control of internal agent components (normative and focus constraints) [Anderson, 1995]. Each of these types has similar components, including a specified active lifetime, activation requirements, and an attachment to some larger knowledge structure. Constraints are organized in a loose hierarchy (Figure 3) abstracting both knowledge and control, and perform different functions ....

[Article contains additional citation context not shown here]

Anderson, John, and Mark Evans, "A Generic Simulation System for Intelligent Agent Designs", Applied Artificial Intelligence 9:5, 1995, pp. 527-562.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC