OPTAIR:     Multi-criteria optimization for air quality
management and emission control

1. Technical objectives

The objective of OPTAIR is to provide tools for the rational design of effective and cost efficient environmental policies for air quality management and emission control. The project results are implemented as a web accessible eConsulting service or, alternatively, as an optional licensed software component: this constitutes a major extension and completely new set of functions of the AirWare system that will greatly enhance its utility and thus market value beyond compliance monitoring and environmental reporting) within the framework of the ongoing EUREKA project E!3266 WEBAIR.

Efficient multi-criteria optimization of control measures that simultaneously considers environmental, technological, and economic criteria adds a completely new dimension and quality of service that is directly addressing growing concerns about the environment and in particular, climate change and the related emission control strategies and policies, going beyond classical air pollutants including CO2 and GHG as covered by the Kyoto protocol. Adding economic assessment and cost-effectiveness as part of the multi-criteria optimization clearly makes the system directly policy relevant, and by demonstrating to be saving money will help to justify its costs to any potential end user.

The overall objective is scientifically sound decision support for environmental management and policy making. The development of the necessary tools and procedures must address several closely related technical challenges:

• Physical and Techno-Economic Realism: The approach is based on state-of-the-art full resolution dynamic, distributed (3D) complex (photochemical) models. These include non-hydrostatic meteorological prognostic models for detailed fields of dynamic meteorological boundary conditions.
  Multi-criteria optimization for non-differentiable, complex systems such as air quality and emission control pose specific methodological problems that include dynamic, distributed, highly non-linear (in particular for ozone and wind entrained dust) systems that defy classical gradient based mathematical programming, but at the same time have considerable computational requirements (approximate ratio of 1/10 real-time on a state-of-the-art dual-processor PC or workstation server system) that precludes any brute forward computational (Monte-Carlo, shotgun) approach to the inverse solving of non-differentiable optimization;
  Technical and performance objective: model validation against air quality observation data according to Council Directive 1999/30/EC, Appendix VIII

• Economic Efficiency: Cost-effectiveness and economic efficiency are important considerations in a policy domain that involves huge costs and benefits. The potential for economic damages but also gains in terms of competitiveness in a system of ultimately global emission control and emission permit trading is very large indeed. Emission control for classical air pollutants (SO2, CO, NOx, PM10/2.5,) is closely related to the control of emissions of CO2 and GHG as covered by the Kyoto protocol. Since these substances are stoichiometrically linked for given fuels, an emission inventory for CO2 can be tested through a bottom up approach involving air quality modelling of more directly measurable pollutants such as SO2 and NOx against air quality monitoring data. This can be used to validate detailed, source specific and bottom- up estimates of CO2 and GHG emissions versus the standard top down macroeconomic approaches, providing an invaluable second, independent estimate.

  Technical and performance objective: computation and display of relative and absolute improvements through optimization expressed as percentage and delta of selected performance criteria compared to the respective baseline scenarios.

• Robustness: Optimization of air quality and emission control is necessarily optimization under considerable uncertainty. The robustness (and thus credibility) of solutions for emission control depends to a considerable degree on the stochastic nature of the driving meteorological conditions, as well as the uncertainties of the emission estimates. One possible approach to find robust and reliable solutions as the basis of sound environmental policies is to jointly optimize ensembles of driving conditions, including alternative IPCC scenario based GCM results downscaled to the level of regional and local urban/industrial air quality problems. Solutions that are pareto-optimal across a range of conditions and GCM scenarios are obviously robust, while economic efficiency can be one of the criteria and objectives directly covered in the multi-criteria optimization approach.

  Technical and performance objective: demonstrated robustness (pareto optimality) of efficient solutions across several meteorological scenarios, including CGM results based on IPCC/SRES A2 and B2.

  To meet these objectives, the project will develop and implement as an operational on-line eConsulting service, an innovative optimization methodology that can combine the full resolution and complexity of the underlying models with sufficient computational efficiency to support an inverse solution to non-differentiable system optimization. This will use:

  • A combination of different models for screening and detailed analysis;
  • Adaptive domain-specific heuristics, genetic algorithms, machine learning;
  • Distributed cluster- and grid computation parallel implementation (see, for example: http://gridengine.sunsource.net).

  OPTAIR builds on the simulation system, models and data bases, available in EUREKA E!3266 WEBAIR and largely developed in FFG PROJECT Nr.808994.