The evaluation of an organization’s environmental performance is an integral part of a corporate environmental management information system. This chapter considers an organization’s environmental impact assessment with respect to a water resource. It investigates formal approaches to the development of temporal monitoring designs for producing data sufficient to perform the assessment. In this study, simple random sampling, stratified random sampling, and designs obtained using greedy search have been investigated with respect to their compatibility with a corporate environmental management information system. All three approaches determine temporal monitoring designs with minimal costs and supply data sufficient for estimation of water quality indicators for a given level of uncertainty. It is shown that monitoring designs obtained using the greedy search approach will outperform other designs when the level of uncertainty in the estimate must be low. If high levels of uncertainty are tolerable, simple random designs become preferable due to their simplicity and effectiveness. The proposed approaches lead to automated procedures which can be easily integrated into a corporate environmental management information system.
Rapidly Learning Bayesian Networks for Complex System Diagnosis: A Reinforcement Learning Directed Greedy Search Approach
