Getting started
The ConstrainedPlanningToolbox is a toolbox for planning under uncertainty in problem domains with resource constraints. The toolbox can be used to formalize and solve planning problems using several built-in planning algorithms. The resulting solution can be evaluated in a simulator that represents a stochastic environment.
The toolbox has been designed based on a simple workflow that consists of three steps: formalize a planning problem with constraints, solve the problem and evaluate the solution. These steps also align with the packages in the source code. A visual representation of the workflow and additional details are provided below. For each step we indicate the relevant input and output.
In the first step a planning problem with constraints is defined based on an application domain. Such a planning problem is formalized as a problem instance.
Output: problem instance
In the second step a planning algorithm can be used to obtain a solution for a given problem instance.
Input: problem instance
Output: solution
The solution computed by a planning algorithm can be evaluated using a simulator, which keeps track of several statistics regarding the performance of the solution in the given domain.
Input: solution
Output: performance statistics
More detailed information about the specific steps can be found on the following pages:
In order to get started with the toolbox, we recommend to read these pages carefully. After that, it is time to start testing based on the examples provided.