De Monfort University
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| Intelligent Agents in Forest Planning System | ||||||||||||||||||||||||||||||||||||||
| By | ||||||||||||||||||||||||||||||||||||||
| Mohd Rasol Abdul Manaf | ||||||||||||||||||||||||||||||||||||||
| Department of Intelligent Computing and Information De Monfort University |
||||||||||||||||||||||||||||||||||||||
| Leicester, UK | ||||||||||||||||||||||||||||||||||||||
| Definition: Intelligent Agents� (IA): | ||||||||||||||||||||||||||||||||||||||
| Franklin and Graesser (1996) reviewed various definitions of the term agent. The most general way to describe the notion is derived as stated below: | ||||||||||||||||||||||||||||||||||||||
| Agents are software entities, which are dedicated to a specific purpose and carry out some set of operations in order to accomplish tasks. | ||||||||||||||||||||||||||||||||||||||
| From a forest planning management perspective this definition is too simple.� Agents taking responsibility in fulfilling the operational goals as well as agents supporting these tasks in a computational way are needed. These thoughts motivate the introduction of two generic meta types of agents in the forest planning domain. �Management agents pursue goals with respect to their environment and their defined action space, whereas their contractors, the service agents solve well specified tasks autonomously.� Management agents need various algorithmical problem solving methods, such as scheduling of tasks with respect to certain conditions, so it makes sense to delegate these kind of tasks to computational agents particularly designed for such purposes. With this task distribution, management agents are able to focus on decision-making problems. | ||||||||||||||||||||||||||||||||||||||
| Management Agents | ||||||||||||||||||||||||||||||||||||||
| Management agents are the central part of a management system and defined �as follows: | ||||||||||||||||||||||||||||||||||||||
| Management agents in forest planning are software entities for meeting operational goals on behalf of human actor or another managerial agent with some degree of independence or autonomy, and in doing so employ some knowledge representation of the user's goals or desires. | ||||||||||||||||||||||||||||||||||||||
| More explicitly: management agents are goal directed, pro-active, take goal responsibility, make decision, and have a model of their environment. They act autonomously, but the actions are constrained by the provided information and models 3 (e.g. from a supervising agent). A management agent needs, therefore, the following information: | ||||||||||||||||||||||||||||||||||||||
| . a goal to pursue | ||||||||||||||||||||||||||||||||||||||
| . its skills and behavior | ||||||||||||||||||||||||||||||||||||||
| model of its environment | ||||||||||||||||||||||||||||||||||||||
| . its role in the agent community (e.g. supervising agent, its sub-agents, collaborating agents) | ||||||||||||||||||||||||||||||||||||||
| . communication and cooperation protocols. | ||||||||||||||||||||||||||||||||||||||
| Since we are in a forest planning domain, where the coordination of joint actions play an important role, a management agent must exhibit the following properties: | ||||||||||||||||||||||||||||||||||||||
| .act in a collaborating manner | ||||||||||||||||||||||||||||||||||||||
| . apply various problem solving strategies | ||||||||||||||||||||||||||||||||||||||
| . communicational abilities | ||||||||||||||||||||||||||||||||||||||
| . methods for solving conflicts among its sub-agents | ||||||||||||||||||||||||||||||||||||||
| . capabilities for goal and model building, which are used by its sub-agents | ||||||||||||||||||||||||||||||||||||||
| This definition was adapted from the IBM definition for intelligent agents (see Franklin and Graesser (1996). | ||||||||||||||||||||||||||||||||||||||
| Gasser (1992) outlined the benefits of using models in DAI Systems.� This is what Martin (1993) calls enterprise modeling, a technique of information engineering.� The previously described planner, controller and operational agents (see section Agents in Forest Planning) have to show all of the above mentioned� attributes, therefore they are management agents. | ||||||||||||||||||||||||||||||||||||||
| Service Agents | ||||||||||||||||||||||||||||||||||||||
| As mentioned earlier in this paper instances of this agent type mainly work on behalf of managerial agents: | ||||||||||||||||||||||||||||||||||||||
| Service agents are contractors of other agents with special computational capabilities without having any representation of their client?s goals. | ||||||||||||||||||||||||||||||||||||||
| Typical tasks for service agents in the domain of forest planning are: | ||||||||||||||||||||||||||||||||||||||
| . computing schedules | ||||||||||||||||||||||||||||||||||||||
| . monitoring the forest planning system for special events, e.g. tree growth projections and treatment scheduling . | ||||||||||||||||||||||||||||||||||||||
| . computing performance data related to forest planning such as estimate forest yield. | ||||||||||||||||||||||||||||||||||||||
| Moreover, service agents are free to split tasks among sub-agents. | ||||||||||||||||||||||||||||||||||||||
| The agent concepts | ||||||||||||||||||||||||||||||||||||||
| In this work, the term agent in its weak notion according to Wooldridge and Jennings (1995) is used.� An agent is an entity in a software-based computer system with the following properties: | ||||||||||||||||||||||||||||||||||||||
| autonomy: agents operate without the direct intervention of humans or others, and have some kind of control over their actions and internal state; | ||||||||||||||||||||||||||||||||||||||
| social ability: agents interact with other agents (and possibly humans) via some kind of agent-communication language; | ||||||||||||||||||||||||||||||||||||||
| reactivity: agents perceive their environment, (which may be the physical world, or the model), and respond in a timely fashion to changes that occur in it; | ||||||||||||||||||||||||||||||||||||||
| pro-activeness: agents do not simply act in response to their environment, they are able to exhibit goal-directed behaviour by taking the initiative. | ||||||||||||||||||||||||||||||||||||||
| Examples: | ||||||||||||||||||||||||||||||||||||||
| SEIDAM: Planning in SEIDAM | ||||||||||||||||||||||||||||||||||||||