Transformational Analogy: A General Framework, Semantics and Complexity
Sarat Chandra Vital Kuchibatla Venltata · 2006
This thesis presents TransUCP, a general framework for transformational analogy. The particulars and working of this framework are further explained with relevant working examples. It is proved, using TransUCP, that transformational analogy does not meet the worst-case complexity scenario of Nebel and Koehler (1995). By generating counterexamples, the framework is also used to prove that their results about plan adaptation being harder than planning from first principles do not apply to it. The implementation details of TransUCP and the results derived from the experimentation are also explained. A synopsis of the TransUCP framework has been accepted for publication at the upcoming European Conference for Case-Based Reasoning (ECCBR-06).