Formulating and solving sequential decision analysis models with continuous variables

This paper presents a new decision analysis approach for modeling decision problems with continuous decision and/or random variables, and applies the approach to a research and development (R&D) planning problem. The approach allows for compact, natural formulation for classes of decision problems that are less appropriately addressed with standard discrete-variable decision analysis methods. Thus it provides a useful alternative analysis approach for problems that are often addressed in practice using simulation risk analysis methods. An illustrative application is presented to energy system R&D planning. The continuous-variable version of this model more directly represents the structure of the decision than a discrete approximation, and the resulting model can be efficiently solved using standard nonlinear optimization methods