Information Technology Reference
In-Depth Information
Algorithm 14.2
Deliberate-Agent algorithm.
Input: static:
environment,
//describing the current world environment
Kb,
//knowledge base
plan
// Plan
Method:
1.
environment
← Update-World-Model(
environment,percept
)
2.
state
← Update-Mental-State(
environment,state
)
3.
plan
← Decision-Making(
state,kb,action
)
4.
environment
← Update-World-Model(
environment,action
)
5. return
action
In the program above, Update-World-Model function generate the abstract
description of current world environment from percept. Update-Mental-State
funcation revises the agent's inner mental state according to the percept
environment. The knowledge base includes the general knowledge and real
knowledge. Agents apply the knowledge to make decision with the Decision-
Making function.
BDI model which is a core of typical deliberative agent can be described with
the following elements:
· a group of beliefs about the world
· current goals
· a plan base which describe how to achieve the goal and revise the belief;
· an intention structure to describe how to achieve the current goals and
revise the current beliefs.
According to BDI ArchitectureRao and Georgeff proposed a simple BDI
interpreter (Rao et al,1992).
Algorithm 14.3
BDI-Interpreter
1. BDI-Interpreter{
2. initialize-state()
3. do
4. options := option-generator(event-queue, B, G, I)
5. selected-options := deliberate(options, B, G, I)
6. update-intentions(selected-options, I)
7. execute(I)
