What problems are solved by the Rapid Recursive® Toolbox?
The Rapid Recursive® Toolbox solves problems known as discrete-time, infinite-horizon Markov decision problems, dynamic programming problems, stochastic control problems or recursive problems. The decision makers in these models must be maximizing their expected discounted reward. Supported Intelligence refers to these problems as “sequential decision problems”.
Does the Rapid Recursive® Toolbox solve models with continuous state variables?
Models with continuous state variables are solved by “discretizing” the variable. Discretization is a standard practice in numerical methods. Discretizing a continuous variable involves converting the variable into a series of points that estimate the continuous variable. For example, an interval from 0 to 1 can be discretized by converting it to the following points 0, 0.1, 0.2, 0.3,…, 0.9, 1.0.
Models with a large number of states are more easily created and solved using the RRvalueiteration or RRpolicyiteration functions and the MATLAB® programming interface than using the Compose Tool.
If the variable is unbounded, you will also need to truncate the variable by placing upper and lower bounds on the variable.
Does the Rapid Recursive® Toolbox solve models with unbounded variables?
Unbounded variables need to be truncated by placing upper and lower bounds on the variable. Truncation a standard practice in numerical methods.