Assessing relocation strategies of urban air quality monitoring stations by GA-based compromise programming

This paper presents a GA-based compromise programming technique for assessi ng the relocation strategy of urban air quality monitoring network with res pect to the multi-objective and multi-pollutant design criteria. While the impact of conservative, quasi-stable, and reactive pollutants are considere d in the design principles via a simulation analysis, cost, effectiveness, and efficiency characteristics are postulated in the optimization process. Therefore, technical coverage for illustrating the needs of siting air qual ity monitoring stations (AQMS) includes both the air quality simulation and optimization modeling analyses in a two-stage analytical framework simulta neously. It starts from determining the spatial interrelationship among tho se candidate sites using various types of air quality simulation models as an integrated means. And the outputs drawn from the simulation models can t hen be used as the required inputs in the compromise programming model in o rder to screen all those siting alternatives that may satisfy the planning goals subject to the essential constraints throughout the multiobjective op timization process. For the illustrating purposes, a series of technical se ttings for finding the optimal relocation scenarios of AQMS were examined i n the case study for the city of Kaohsiung in South Taiwan where the long-t erm violations of official standards of ozone and particulates turn out to be critical. It not only expresses the ideas of relocation strategy but als o indicates how to utilize those alternatives in the decision-making proces s for improving the functionality of air quality monitoring in the urban en vironment. Experience gained in this study clearly indicates that the more the number of pollutants and objectives considered simultaneously, the high er the number of candidate sites to be selected in the relocation strategy. (C) 2001 Elsevier Science Ltd. All rights reserved.