“…The problem of optimal sensor placement for state estimation has attracted a large body of practical and theoretical research. On the practical side, sensor placement has been investigated with respect to state estimation in complex dynamical systems like electric power grids (Cho et al., 2001; Dua et al., 2008), biochemical reaction networks (Liu et al., 2013), structural systems (Safizadeh & Darus, 2010; van der Linden et al., 2011; Weickgenannt et al., 2011), weather systems (Guestrin et al., 2005), and contaminant monitoring in soil and groundwater (Castello et al., 2010; Reed et al., 2000). While application areas are diverse, these studies share the common goal of designing sensor networks that can optimally infer system states from limited observed data—for instance, estimating line voltages in an electrical grid from data collected at generating stations.…”