Minimizing Age-of-Information with Throughput Requirements in Multi-Path Network Communication
share
We consider the scenario where a sender periodically sends a batch of data to a receiver over a multi-hop network using multiple paths. Our objective is to minimize peak/average Age-of-Information (AoI) subject to both a minimum and a maximum throughput requirement, by jointly optimizing throughput and multi-path routing strategy. The consideration of batch generation and multi-path communication differentiate us from existing studies. First, we show that our AoI minimization problems are NP-hard, but only in the weak sense as we develop an optimal algorithm with a pseudo-polynomial time complexity. Next we show that minimizing AoI and minimizing maximum delay for a batch of data are "largely" compatible. In particular, we show that any optimal solution of minimizing maximum delay is an approximate solution of minimizing AoI with bounded optimality loss. We leverage this understanding to develop a framework to adapt any polynomial-time α-approximation algorithm of the existing maximum delay minimization problem to develop a polynomial-time algorithm for our AoI minimization problems with an approximation ratio of α+β, where β is a constant determined by input minimum and maximum throughput requirement. The framework suggests a new avenue for designing approximation algorithms for minimizing AoI in multi-path communications. Extensive simulations that compare our approximation framework and our optimal algorithm over various network topologies validate the effectiveness of our proposed approaches.
READ FULL TEXT