Information Technology Reference
In-Depth Information
f
=
γ(
false
,
s
)
true
.
false
Then we have:
<goal
[
extract
,
fail
,
result
, <
goal
2
[
n
]>]> = <
t
,
ff
,
true
,
false
,
EM
,
false
,
∅:
R
> where
Let
e
=
extract
,
fail
,
result
, <
goal
2
[
n
]>
true
= γ(
e:EM
)<
goal'
[
e
]>=<
true
>.
true
,
e
:
R
,
true
false =
γ(
e:EM
)
< goal'
[
e
]>=<
false
>.
false
,
false
ff =
γ(
false
,
e
)
e≠
true
∧
type
(
e
)
≠R
.
false
t
=
γ(
true
,
e
)
type
(
e
)
≠R
.
true
From the depicted process we can see that the conversion from logic specification to operational
specification can be done using a small number of rules and we can see that the process can be automated
with a parsing program. Let's highlight that this approach allows a large complex system to be design
modularly. A complex logic specification can be decomposed into simpler specifications and converted
into γ programs separately. With the conceptual model of γ-Calculus, it is very easy to combine verifica-
tion programs (in the form of solutions) into a compound module by using a compassing solution.
iMPle Ment ation
The above course material updating system has been implemented and simulated using the Message
Passing Interface (MPI). This section details the design of the e-learning agent system, the design and
architecture of the agents themselves, and explanation of MPI and how MPI facilitates the simulation.
The System Architecture
In order to provide a comprehensive e-learning environment, the system needs to account for:
•
Changes in the curriculum:
Adding or removing course content
•
Solving problems that occur:
Reporting broken Web links, and fixing them
•
Customizing content presentation:
Displaying course content based on the individual user's
needs
•
Matching student models:
Delivering course content based on the specific student's require-
ments
Using current technology, the implementation of a distributed system, that satisfies these require-
ments, results in a complex, difficult to administer, and cost prohibitive solution. The complexity of
the system is derived from individual components competing for resources and engaging in complex
communications in order to deliver data and satisfy their individual goals.
For this project, there are two types of agents to discuss; interface agents and information agents.
Interface Agents:
Interface agents are designed to interact directly with the user or other agents. In
the e-learning system developed for this project, they are referred to as control agents. They provide a
mechanism for accepting commands from a user, processing the command, and delivering the resulting
Search WWH ::
Custom Search