+ All Categories
Home > Documents > 15 August, 2005IEEE IRI 20051 Web Based Expert System for Class Schedule Planning using JESS Ken Ho...

15 August, 2005IEEE IRI 20051 Web Based Expert System for Class Schedule Planning using JESS Ken Ho...

Date post: 04-Jan-2016
Category:
Upload: barnaby-henry
View: 235 times
Download: 0 times
Share this document with a friend
Popular Tags:
18
15 August, 2005 IEEE IRI 2005 1 Web Based Expert System Web Based Expert System for Class Schedule for Class Schedule Planning using JESS Planning using JESS Ken Ho Hewlett Packard Company Meiliu Lu Department of Computer Science California State University, Sacramento 15 August, 2005
Transcript

15 August, 2005 IEEE IRI 2005 1

Web Based Expert System for Class Web Based Expert System for Class Schedule Planning using JESSSchedule Planning using JESS

Ken HoHewlett Packard Company

Meiliu LuDepartment of Computer Science

California State University, Sacramento

15 August, 2005

15 August, 2005 IEEE IRI 2005 2

AgendaAgenda

Motivation Related Works Paper Contributions System Architecture & Design Short Demo Conclusion Future Works Questions

15 August, 2005 IEEE IRI 2005 3

MotivationsMotivations

Class schedule planning mistake often causes graduation delay

Demands for a convenient, intelligent system to give expert advises on class schedule planning decisions

Current solutions have one or more deficiencies:– Not web-enabled– Few user-defined inputs– Require rewriting the tool on requirement

changes

15 August, 2005 IEEE IRI 2005 4

Related WorksRelated Works OASIS advising assistance system from UC Davis

– A command-line tool that reports discrepancies about a academic plan

Degree Audit Tool from Chico State University.– A web-based tool that shows a countdown of gradation

requirement Proposed Expert Systems on academic planning [1, 2, 3, 4]

15 August, 2005 IEEE IRI 2005 5

Class Schedule Planner (CSP) Class Schedule Planner (CSP) ContributionsContributions Considers various user-defined parameters

Concentration (ex. A.I., Computer Architecture…) Part time or full time student Courses they already completed Preferred times and days to attend classes Courses that they like to take Courses that they want to avoid

Allows dynamic management of knowledge in real time using web interfaces

15 August, 2005 IEEE IRI 2005 6

CSP OverviewCSP Overview

CSP – Provides a set of web forms to collect inputs from students– Uses encapsulated expert knowledge to provide scheduling

advises– Collects current class data from school web sites in real time– Provides web forms for system administrators to change

requirements System Users

– Students• Request suitable schedules by answering a set of questions

– Administrators• Update degree requirement and class prerequisites• Update the information URL (current semester class schedule)

15 August, 2005 IEEE IRI 2005 7

CSP Overall ArchitectureCSP Overall ArchitectureClass Schedule Planner (CSP)

data file (xml)

static rules

Jess

XMLtranslator

dynamic facts

CSPController

Admin web browser

User web browser

Jess output

Internet

defaultdegree requirement

&course prerequisites

User parameters

class schedule(s)(up to 4 terms)

updatedegree requirement

andcourse prerequisites

previous user data

15 August, 2005 IEEE IRI 2005 8

What is JESS?What is JESS?

An Open-source Expert System Shell for the Java Platform– Similar to the language defined by CLIPS

expert system shell Supports the development of rule-based

expert system Provides Java interfaces to store and map

expert system rules and facts Good for web-based expert system with

changing rules and facts

15 August, 2005 IEEE IRI 2005 9

Class Schedule Planner (CSP)

data file (xml)

XMLtranslator

CSPController

Internet

User web browser

static rules

Jess

Student InteractionStudent Interaction

Jess output

User parameters

class schedule(s)

requirement data

dynamic factsdynamic rules

class schedule(s)

User parameters

User parameters

15 August, 2005 IEEE IRI 2005 10

Goal Driven Approach with JESSGoal Driven Approach with JESS

Stage 1– Provides 1st set of recommended classes

based on the student inputs Stage 2

– Provides 2nd set of recommended classes based on the prerequisites of the 1st set of recommended classes

Stage 3– Provides 3rd set of recommended classes

based on the random selection of remaining classes

15 August, 2005 IEEE IRI 2005 11

Class Schedule Planner (CSP)

data file (xml)

XMLtranslator

CSPController

Internet

Admin web browser

Administrator InteractionAdministrator Interaction

Updatedegree requirement

&Course prerequisites

updatedegree requirement

&course prerequisites

updatedegree requirement

andcourse prerequisites

updatedegree requirement

andcourse prerequisites

15 August, 2005 IEEE IRI 2005 12

Dynamic Change of RulesDynamic Change of Rules

The following is an example of the Core requirement rule: (defrule satisfy_core(course_taken (name "CSC 201"))(course_taken (name "CSC 204"))(course_taken (name "CSC 205"))(course_taken (name "CSC 206"))(course_taken (name "CSC 209")) (not (core_passed yes)) => (assert (core_passed yes)))

(defrule satisfy_core(course_taken (name "CSC 201"))(course_taken (name "CSC 204"))(course_taken (name "CSC 205"))(course_taken (name "CSC 206"))(course_taken (name "CSC 209")) (not (core_passed yes)) => (assert (core_passed yes)))

(defrule satisfy_core(course_taken (name "CSC 201"))(course_taken (name "CSC 204"))(course_taken (name "CSC 205"))(course_taken (name "CSC 206"))(course_taken (name "CSC 207"))(course_taken (name "CSC 209")) (not (core_passed yes)) => (assert (core_passed yes)))

15 August, 2005 IEEE IRI 2005 13

Demo: Make SchedulesDemo: Make Schedules

15 August, 2005 IEEE IRI 2005 14

ConclusionConclusion

We integrate three key elements in this paper– Understanding problem areas in class

schedule planning domain– Integrating JESS, Java, XML and

popular web technologies for easy web-based deployment

– Providing methodologies for dynamic knowledge management

15 August, 2005 IEEE IRI 2005 15

Future worksFuture works

Making the degree requirement specification more scaleable

Data warehousing CSP users data for empirical evaluation and suggestions for new student users

Connect to the student information database directly so that part of student user input like courses taken can be automated

Port CSP server to Linux/Unix platforms – We are currently porting CSP to the CSUS CS department’s

Linux server

15 August, 2005 IEEE IRI 2005 16

ReferenceReference

[1] Tsutsui, S. et al, “Class scheduling by neurocomputing: a comparison with the expert system approach”, Systems, Man and Cybernetics, 1990. Conference Proceedings, IEEE International Conference on, Vol., Iss., 4-7 Nov 1990, Pages:227-232

[2] Yoshikawa, M et al, “A constraint-based high school scheduling system”, Expert, IEEE [see also IEEE Intelligent Systems], Vol.11, Iss.1, Feb 1996, Pages: 63-72

[3] McDonald, G et al, “An expert system solution to a constraint satisfaction problem in academic administration”, Applied Computing, 1990., Proceedings of the 1990 Symposium on, Vol., Iss., 5-6, Apr 1990 Pages:161-163

[4 ] Frank, J.L et al, “Mentor-I: an expert database system for student guidance”, Expert, IEEE [see also IEEE Intelligent Systems], Vol.3, Iss.2, Summer 1988, Pages: 40-46

15 August, 2005 IEEE IRI 2005 17

QuestionsQuestions

Thank You!Thank You!


Recommended