|
Probabilistic inference engines: a foundation
for reasoning with uncertainty with application to environmental modeling
|
Advanced Software Solutions in S&T (ASSIST) |
|
Proposal number: |
23213 |
|
Focal area: |
Information Society |
|
Total funding: |
R 2,725,000 |
|
Funding year 1: |
R 925,000 |
|
Funding year 2: |
R 900,000 |
|
Funding year 3: |
R 900,000 |
|
The study of environmental pollution is a subject of rapid
growth globally. For example, predicting the dispersion of
air-borne pollutants arising from various sources is important for
planning, regulation and health and safety management. In this
context, atmospheric dispersion modeling may be used to predict
concentrations of dispersed pollutants given knowledge of the
source properties: where and how intense the sources are. The
"inverse" problem involves the opposite process; using
dispersed emission measurements to infer unknown source
properties, viz. source locations and strengths, and indeed how
many they are. This problem may be of interest in situations such
as detecting leaks in chemical plants. It is a challenging problem
that requires considerable innovation in the development of tools
for quantitative diagnosis of causes (emission sources) given
observations of their effects (dispersed emission concentrations
measured at a finite number of spatial locations). This project
seeks to develop software tools to deal with this important
problem in air quality studies. |
|
Project Coordinator: |
Dr Sibusiso Sibisi
Phone: +27 (0)21 710 2142
Email: sibisis@bremner.uct.ac.za |
Public reports / Newsletters:
| Back
to projects table | Top of
page |
|
|