Optimal experiment design in a filtering context with application to sampled network data

TL;DR

This paper develops a method for optimal experiment design in filtering multiple random walks, focusing on network flow tracking with sampled data, and formulates the problem as a second order cone program.

Contribution

It introduces a steady state E-optimal design criterion for filtering, leading to a second order cone program, and applies it to network flow volume tracking with sampling rate control.

Findings

Optimal design outperforms naive strategies in simulations.

The methodology is applicable to steady state design for state space models.

The problem reduces to a second order cone program for efficient computation.

Abstract

We examine the problem of optimal design in the context of filtering multiple random walks. Specifically, we define the steady state E-optimal design criterion and show that the underlying optimization problem leads to a second order cone program. The developed methodology is applied to tracking network flow volumes using sampled data, where the design variable corresponds to controlling the sampling rate. The optimal design is numerically compared to a myopic and a naive strategy. Finally, we relate our work to the general problem of steady state optimal design for state space models.