You are here:
Genetic Algorithms in the Environmentally Conscious Design of a Softwood Kraft Pulp Bleach PlantEPA Grant Number: U915199
Title: Genetic Algorithms in the Environmentally Conscious Design of a Softwood Kraft Pulp Bleach Plant
Investigators: Clayton, John M.
Institution: Georgia Institute of Technology - Main Campus
EPA Project Officer: Jones, Brandon
Project Period: September 1, 1997 through August 1, 2000
Project Amount: $102,000
RFA: STAR Graduate Fellowships (1997) RFA Text | Recipients Lists
Research Category: Fellowship - Environmental Engineering , Academic Fellowships , Engineering and Environmental Chemistry
The objective of this research project is to devise a general method for solving environmentally conscious industrial design problems that contain detailed, fundamental environmental process modeling. The general method will be based on a genetic algorithm (GA) coupled with cluster analysis.
A GA will be used to solve this design problem. GAs are a class of sequential discrete random sampling search techniques that use the best performing old solutions to guide the search for newer, better solutions. Near-optimal solutions evolve over the course of the algorithm. GAs are known to accept nonlinear, discontinuous, or even differential objective and constraint functions. The results of the GA will be a set of near-optimal solutions to the design problem. Many of these solutions will have similar design parameter values, suggesting ranges of values that lead to near-optimal system performance. These ranges will be summarized by nearest neighbor cluster analysis on the set of solutions and by determining the multivariate mean and prediction interval for each cluster. Knowledge of the ranges of design parameter values that lead to near optimality can communicate the design freedom associated with optimal solutions. These characterizations also can serve as design spaces for more refined design problems.