Track:
All Papers
Downloads:
Abstract:
Many software applications consist of a number of interdependent steps and are executed under the supervision of a human administrator or operator. The administrator responds to error conditions and checks to see that the dependencies between steps has been satisfied before allowing the application to continue. AI planning systems, because they have the capability of responding to a dynamically changing environment by replanning, are an excellent candidate for automating such applications. We report on a project that uses AI planning technology as part of the solution in automating a corporate data mining application.