Combinatorics of Crossing Networks
Geoff A. Latham · Defense Technical Information Center (DTIC) · 2000
In scenarios such as software agents sharing time slots or blocks in different physical CPUs or memories, communications agents accessing channels or bandwidth on shared links or bearers and economic agents trading in commodities in financial markets, the agents in question can exchange resources among themselves for mutual benefit. This report develops a combinatorial probability model to describe the like-for-like exchange of resources between multiple unbiased agents. The expected exchange rates are computed for individual agents, syndicates of agents and the collection of all agents. These performance benchmarks provide intuition and understanding for the performance of real crossing networks. A number of combinatorial identities are also produced as a consequence of the modelling and analysis.