Article Contents
Article Contents

# Dynamic observers for unknown populations

• * Corresponding author
• Dynamic observers are considered in the context of structured-population modeling and management. Roughly, observers combine a known measured variable of some process with a model of that process to asymptotically reconstruct the unknown state variable of the model. We investigate the potential use of observers for reconstructing population distributions described by density-independent (linear) models and a class of density-dependent (nonlinear) models. In both the density-dependent and -independent cases, we show, in several ecologically reasonable circumstances, that there is a natural, optimal construction of these observers. Further, we describe the robustness these observers exhibit with respect to disturbances and uncertainty in measurement.

Mathematics Subject Classification: Primary: 39A30, 92D40, 93B51, 93B52, 93C05, 93C10, 93D09.

 Citation:

• Figure 1.1.  Illustration of a dynamic observer. The dashed lines in the right-hand figure at computed using the measured variable $y(t)$

Figure 5.1.  Numerical simulations from Example 5.1. Abundances from the model (LO) are plotted against time. The solid line and dotted lines are the abundance of the eighth stage-class $x_8(t)$ of the (forced) cheetah population, and its corresponding observer state $z_8(t)$, respectively. The dashed and dashed-dotted lines are $\| x(t)\|_1$ and $\| z(t)\|_1$, the total population and its observer estimate, respectively

Figure 5.2.  Numerical simulations from Example 5.2. In both panels, the solid, dashed, dashed-dotted and dotted lines denote the total population $\| x(t)\|_1$, the corresponding estimate $\| z(t)\|_1$, the error $\| x(t) - z(t) \|_1$ and the unforced equilibrium $\| x^*\|_1$, respectively, each plotted against time $t$. In (a), no forcing terms are present, so $d = 0$ and $v = 0$, and the error converges to zero. In (b), the forcing and measurement error terms are nonzero, described in the main text

Figure 5.3.  Numerical simulations from Example 5.3. The solid line denotes $\|x(t)\|_1$, and the dashed, dashed-dotted, and dotted lines denote the errors $\| x(t) - z(t)\|_1$ for increasing measurement error. See the main text

Figures(4)