We present a novel method for building ABSTRIPS style abstraction hier
archies in planning. The aim of this method is to minimize search by l
imiting backtracking both between abstraction levels and within an abs
traction level. Previous approaches for building ABSTRIPS style abstra
ctions have determined the criticality of operator preconditions by re
asoning about plans directly. Here, we adopt a simpler and faster appr
oach where we use numerical simulation of the planning process, We dev
elop a simple but powerful theory to demonstrate the theoretical advan
tages of our approach, We use this theory to identify some simple prop
erties lacking in previous approaches but possessed by our method, We
demonstrate the empirical advantages of our approach by a set of four
benchmark experiments using the ABTWEAK system, We compare the quality
of the abstraction hierarchies generated with those built by tie ALPI
NE and HIGHPOINT algorithms.