Author

Rui Zhang

Date of Award

5-2005

Level of Access

Open-Access Thesis

Degree Name

Master of Science (MS)

Department

Spatial Information Science and Engineering

Advisor

Michael F. Worboys

Second Committee Member

Max J. Egenhofer

Third Committee Member

Kathleen E. Hornsby

Abstract

Modeling the dynamic aspect, or change, of geographic phenomena is essential to explain the evolution of geographic entities and predict their future. Event-based modelling, describing the occurrences rather than states of geographic phenomena, gives an explicit treatment of such change, but currently does not have the support of the mechanisms to enable the shifts among different granularities of events. To account for different tasks, a hierarchical representation of the event space at different granularities is needed.

This thesis presents an event-based model; a general framework for representing events based on precondition and postcondition using Allen's temporal interval logic. It captures not only the changes to the objects, but also some contextual information that is necessary for the occurrence of events. Analogous to objects, events have types and instances, and two abstraction processes in the object-oriented paradigm, generalization and aggregation, are applied to events. Event-event relations a.re investigated through their preconditions and post,conditions. Our representation of relationships between events is based on two relations between events, f-sequences and f-transitions. These relationships play an important role in describing the structure of a component event in the event partonomy, and therefore provide a mechanism to construct the event partonomy automatically. This research constructs an algorithm to generate the part-whole hierarchy for events, which supports multiple representations of events and enables shifts among them. To illustrate the process of constructing the event partonomy, we give a case study of a car accident scenario.

Share