Selective Encoding Policies for Maximizing Information Freshness

An information source generates independent and identically distributed status update messages from an observed random phenomenon which takes n distinct values based on a given probability mass function (PMF). These packets are encoded at the transmitter node to be sent receiver wants track variable with as little age possible. The implements selective k encoding policy such that rather than all possible realizations, encodes most probable realizations. We consider three different policies regarding remaining n-k less realizations: highest disregards whenever realization occurs; randomized sends realizations certain further inform expense of longer codewords for selected realizations; empty symbol designated when one occurs. For these schemes, we find average determine age-optimal real codeword lengths, including length in case latter scheme, is minimized. Through numerical evaluations arbitrary PMFs, show result lower every realization, corresponding values. Since focus real-valued lengths this paper, resulting value obtained each studied here serves bound what can attained by integer-valued case.