Prioritising Behaviours in Qualitative Simulation

Abstract

We present a simple but potentially extremely powerful technique for associating degrees of commitment to the possible behaviours produced by qualitative simulation algorithms. The 'distance' between the predicted successor states and those determined from local propagation through the constraints is used to determine the most likely state and hence prioritise the behaviours. The method forms a basis for a technique for performing progressive reasoning to cope with realtime constraints. In this manner, the most likely solution is generated first and the subsequent behaviours are only generated if higher priority behaviours fail. Ultimately all possible behaviours are generated, as in current algorithms, and the soundness of qualitative predictions is guaranteed. The approach is applicable to any non-constructive simulation algorithm (e.g. QSIM and its derivatives) when the landmarks are represented by known real numbers or semiquantitative information is utilised.