Many investment decisions involve multiple decision stages that have termination options. Information is revealed over time, and the ability to modify behaviour leads to a decision “tree”. Once uncertain outcomes and a changing environment over time are factored in, solving these types of problems becomes very complex. Stochastic Dynamic Programming provides a solution method that will solve these problems.
These models can be used in applications such as mining or oil exploration where testing can reveal better information at some cost.
A problem with real-world complexity is likely to involve some or all of the following aspects:
Stochastic Dynamic Programming is a very efficient method for solving large decision trees in a short period of time. At each combination of decision nexus and possible future state the SDP solver tells us the optimal decision to make.
Our custom-written solver creates the optimal solution for this type of solution in a very short period of time. As well as the optimal “roadmap” of decisions, outputs include such things as expected return, and expected cash flow.