Scalable Parameter Management using Casebased Reasoning for Cognitive Radio Applications
Cognitive radios have applied various forms of artificial intelligence (AI) to wireless systems in order to solve the complex problems presented by proper link management, network traffic balance, and system efficiency. Casebased reasoning (CBR) has seen attention as a prospective avenue for storing and organizing past information in order to allow the cognitive engine to learn from previous experience. CBR uses past information and observed outcomes to form empirical relationships that may be difficult to model apriori. As wireless systems become more complex and more tightly time constrained, scalability becomes an apparent concern to store large amounts of information over multiple dimensions. This thesis presents a renewed look at an abstract application of CBR to CR. By appropriately designing a case structure with useful information both to the cognitive entity as well as the underlying similarity relationships between cases, an accurate problem description can be developed and indexed. By separating the components of a case from the parameters that are meaningful to similarity, the situation can be quickly identified and queried given proper design. A data structure with this in mind is presented that orders cases in terms of general placement in Euclidean space, but does not require the discrete calculation of distance between the query case and all cases stored. By grouping possible similarity dimension values into distinct partitions called "similarity buckets", a data structure is developed with constant (O(1)) access time, which is an improvement of several orders of magnitude over traditional linear approaches (O(n)).