Optimized selection of wireless network topologies and components via efficient pruning of feasible paths

We address the design space exploration of wireless networks to jointly select topology and component sizing. We formulate the exploration problem as an optimized mapping problem, where network elements are associated with components from pre-defined libraries to minimize a cost function under correctness guarantees. We express a rich set of system requirements as mixed integer linear constraints over path variables, denoting the presence or absence of paths between network nodes, and propose an algorithm for efficient, compact encoding of feasible paths that can reduce by orders of magnitude the complexity of the optimization problem. We incorporate our methods in a system-level design space exploration toolbox and evaluate their effectiveness on design examples from data collection and localization networks.