Cookies

We use cookies to ensure that we give you the best experience on our website. By continuing to browse this repository, you give consent for essential cookies to be used. You can read more about our Privacy and Cookie Policy.


Durham e-Theses
You are in:

Using Distributed Agents to Create University Course Timetables Addressing Essential & Desirable Constraints and Fair Allocation of Resources

WANGMAETEEKUL, PENNEE (2011) Using Distributed Agents to Create University Course Timetables Addressing Essential & Desirable Constraints and Fair Allocation of Resources. Doctoral thesis, Durham University.

[img]
Preview
PDF (Wangmaeteekul PhD Thesis) - Accepted Version
2194Kb

Abstract

In this study, the University Course Timetabling Problem (UCTP) has been investigated. This is a form of Constraint Satisfaction Problem (CSP) and belongs to the NP-complete class. The nature of a such problem is highly descriptive, a solution therefore involves combining many aspects of the problem. Although various timetabling algorithms have been continuously developed for nearly half a century, a gap still exists between the theoretical and practical aspects of university timetabling.

This research is aimed to narrow the gap. We created an agent-based model for solving the university course timetabling problem, where this model not only considers a set of essential constraints upon the teaching activities, but also a set of desirable constraints that correspond to real-world needs. The model also seeks to provide fair allocation of resources. The capabilities of agents are harnessed for the activities of decision making, collaboration, coordination and negotiation by embedding them within the protocol designs. The resulting set of university course timetables involve the participation of every element in the system, with each agent taking responsibility for organising of its own course timetable, cooperating together to resolve problems. There are two types of agents in the model; these are Year-Programme Agent and Rooms Agent. In this study, we have used four different principles for organising the interaction between the agents: First-In-First-Out & Sequential (FIFOSeq), First-In-First-Out & Interleaved (FIFOInt), Round-Robin & Sequential (RRSeq) and Round-Robin & Interleaved (RRInt). The problem formulation and data instances of the third track of the Second International Timetabling Competition (ITC-2007) have been used as benchmarks for validating these implemented timetables. The validated results not only compare the four principles with each other; but also compare them with other timetabling techniques used for ITC-2007.

The four different principles were able to successfully schedule all lectures in different periods, with no instances of two lectures occupying the same room at the same time. The lectures belonging to the same curriculum or taught by the same teacher do not conflict. Every lecture has been assigned a teacher before scheduling. The capacity of every assigned room is greater than, or equal to, the number of students in that course. The lectures of each course have been spread across the minimum number of working days with more than 98 percent success, and for more than 75 percent of the lectures under the same curriculum, it has been possible to avoid isolated deliveries. We conclude that the RRInt principle gives the most consistent likelihood of ensuring that each YPA in the system gets the best and fairest chance to obtain its resources.

Item Type:Thesis (Doctoral)
Award:Doctor of Philosophy
Keywords:multi agent system, university course timetabling problem, modeling complex system, constraint satisfaction problem, resource allocation, essential constraints, desirable constraints,fairness of resource allocation, collaboration, coordination and negotiation
Faculty and Department:Faculty of Science > Engineering and Computing Science, School of
Thesis Date:2011
Copyright:Copyright of this thesis is held by the author
Deposited On:31 May 2012 12:43

Social bookmarking: del.icio.usConnoteaBibSonomyCiteULikeFacebookTwitter