Solution Border Estimation for Efficient Analysis of Plan Space

Problem Space Analysis allows a planner to find patterns within the characteristics of problem scenarios that may be encountered in the near future. One component of such analysis, the problem-space map (PS Map), identifies groups problems that a particular solution can resolve. An ideal PS Map allows a system to preemptively plan for future scenarios, but is generally prohibitively expensive to compute. This talk will give an overview of problem space analysis, discuss the Solution Border Estimation technique, and present initial results.