A Novel Application of Evolutionary Computing in Process Systems Engineering
- Cite this paper as:
- Carballido J.A., Ponzoni I., Brignole N.B. (2005) A Novel Application of Evolutionary Computing in Process Systems Engineering. In: Raidl G.R., Gottlieb J. (eds) Evolutionary Computation in Combinatorial Optimization. EvoCOP 2005. Lecture Notes in Computer Science, vol 3448. Springer, Berlin, Heidelberg
In this article we present a Multi-Objective Genetic Algorithm for Initialization (MOGAI) that finds a starting sensor configuration for Observability Analysis (OA), this study being a crucial stage in the design and revamp of process-plant instrumentation. The MOGAI is a binary-coded genetic algorithm with a three-objective fitness function based on cost, reliability and observability metrics. MOGAI’s special features are: dynamic adaptive bit-flip mutation and guided generation of the initial population, both giving a special treatment to non-feasible individuals, and an adaptive genotypic convergence criterion to stop the algorithm. The algorithmic behavior was evaluated through the analysis of the mathematical model that represents an ammonia synthesis plant. Its efficacy was assessed by comparing the performance of the OA algorithm with and without MOGAI initialization. The genetic algorithm proved to be advantageous because it led to a significant reduction in the number of iterations required by the OA algorithm.
KeywordsCombinatorial Optimization Problem PSE Process-Plant Instrumentation Design Multi-Objective Genetic Algorithm Observability Analysis
Unable to display preview. Download preview PDF.