The value for decision making of high-quality post- solution analysis of decision-supporting models can hardly be overestimated. Candle-lighting analysis (CLA) takes this notion very seriously and brings to the table a series of computational approaches for gaining insight into decision problems by performing heuristic mapping of decision surfaces for management science models. The principal aims of this paper are to present, illustrate, and demonstrate the value, and extensive range, of candle-lighting analysis (CLA) ideas in the context of a new decision support system, CLAP- NT. In addition to illustrating how heuristic decision surface maps may be used for post-evaluation analysis, we present preliminary results comparing the efficacy of a genetic algorithm and a random search for mapping decision surfaces.