Finite-horizon approximation conjecture for extremum-seeking systems

Let x(t)x(t) be a trajectory of systems and, and let xˉ(t)\bar{x}(t) be a trajectory of the averaged approximation

xˉ˙=f(xˉ,t)+ϵg2ne(xˉ,t)(αω)2ne2m+1Bnekα2no+12m+1ω(2no+12m+11)An0gno(xˉ,t)gnoT(xˉ,t)24no+1(Vxˉ)T,\dot{\bar{x}} = f(\bar{x},t) + \epsilon g_{2n_e}(\bar{x},t)\left ( \alpha \omega \right )^{\frac{2n_e}{2m+1}}B_{n_e} - k\alpha^{\frac{2n_o+1}{2m+1}}\omega^{\left ( \frac{2n_o+1}{2m+1}-1 \right )} A_{n_0} \frac{g_{n_o}(\bar{x},t)g^T_{n_o}(\bar{x},t)}{2^{4n_o+1}} \left (\frac{\partial V}{\partial \bar{x}}\right )^T,

where

An0=l=0no(2no+1l)2,Bne=122ne(2nene).A_{n_0} = \sum_{l=0}^{n_o}\binom{2n_o+1}{l}^2, \qquad B_{n_e} = \frac{1}{2^{2n_e}}\binom{2n_e}{n_e}.

Finite-horizon approximation conjecture. For any δ>0\delta>0, any compact set KRnK \subset \mathbb{R}^n, and any t0,TR0t_0,T \in \mathbb{R}_{\geq 0}, there exists ϵ(δ,K,T)>0\epsilon^\star(\delta,K,T)>0 such that, for all ϵ<ϵ|\epsilon|<\epsilon^\star, there exists ω\omega^\star such that for each ω>ω\omega>\omega^\star,

maxt[t0,t0+T]x(t)xˉ(t)<δ.\max_{t\in [t_0,t_0+T]}\|x(t)-\bar{x}(t)\|<\delta.

This conjecture asserts uniform approximation of the original and averaged trajectories on arbitrary finite time intervals under sufficiently small perturbation amplitude and sufficiently high excitation frequency. The supplied text does not state whether the claim has been proved or remains open.

Sources & referencesView supporting material

Primary source

Alexander Scheinker and David Scheinker, “Extremum Seeking for Stabilization of Systems Not Affine in Control”, arXiv:1608.04587 (2016).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.