Simulations in wireless sensor and ad hoc networks: matching and advancing models, metrics, and solutions

The objective of this article is to give advice for carrying out a proper and effective simulation activity for protocol design. It challenges some of the existing criticisms of simulation practices that emphasized validation aspects. This article advocates the use of simple models, matching assumptions and metrics in the problem statement and simulation to provide a basic "proof of concept," and comparison with truly competing solutions, which is possible only after a thorough and critical literature review. Then the complexity of the models can be increased (one parameter at a time), revising the algorithms themselves by adapting them to new assumptions, metrics, and the corresponding simulation environment. Selected independent variables should explain performance under a wide range of scenarios.