The discrete resource allocation problem (RAP) examines the problem of allocation of discrete and indivisible scarce resources in a multiple agent environment from the point of view of a coordinator. The basis for resource allocation is preference information of each agent regarding the discrete resources. Such preference information is obtained by the coordinator through a sequential experimental process that entails interrogating each of the agents about their preference profiles. In this environment, where information is incomplete and has finite cost, two kinds of experimental schemes are investigated : one consists of pairwise comparisons between resource bundles about which there is incomplete a priori preference information, and the other consists of experiments that rank resource bundles about which there is incomplete a priori preference information. The signals that result from each such experiment are used by the coordinator to systematically eliminate all except the one preference ordering that describes the agent’s preference profile. The paper carries out the investigation in terms of entropy, a general nonparametric measure.