Our work focuses upon development of techniques for choosing among a set of altematives in the presence of incomplete information and varying costs of acquiring information. In our approach, we model the cost and utility of various alternatives using parameterized statistical models. By applying techniques from an area of statistics called parameter estimation, statistical models can be inferred from data regarding the utility and information cost of each of the various alternatives. These statistical models can then be used to estimate the utility and cost of acquiring additional information and the utility of selecting specific alternatives from the possible choices at hand. We apply these techniques to adaptive problem-solving, a technique in which a system automatically tunes various control parameters on a performance element to improve performance in a given domain. We present empirical results comparing the effectiveness of these techniques on speedup learning from a real-world NASA scheduling domain and schedule quality data from the same real-world NASA scheduling domain.