Particle Swarm Optimization Enabled Parametric Mapping for Channel Model Substitution

Channel model substitution (CMS) is a technique that aims to replace a computationally challenging channel model with a simpler substitute. This technique is powerful for rapid adaptive signal processing and closed-form performance analytics. The parametric mapping between an original channel model and its substitute determines the utility of CMS. In the past decades, the moment matching criterion has dominated for conducting parametric mapping, which, however, is heuristic and has been proven non-optimal. In this paper, we propose to utilize particle swarm optimization (PSO) to obtain optimal parametric mapping relations for a general CMS problem, regardless of the distributional forms of the original channel model and the substitute. Taking the CMS techniques for the lognormal shadowed channel model as examples, simulation results show that the PSO enabled parametric mapping approach is capable of converging to the global optima under diverse system configurations, making CMS computationally feasible.