Abstract
Large-scale engineering projects often entail large financial and environmental risks because of their size and long life. Decisions usually have to be made long before the risks can be appreciated. Decision theory should be helpful in this but utility curves are needed for its application and these are almost impossible to obtain. The Kelly criterion, which involves maximizing the expected value of the logarithm of wealth, offers an approach to minimizing financial risk; and Shannon's measure of information offers a criterion for the design of networks for environmental monitoring for early detection of hazards whose characteristics may not be known in advance. Both measures implicitly involve logarithmic utility functions.