“微分模型的全局可辨识性”的勘误

IF 4.3 3区 材料科学 Q1 ENGINEERING, ELECTRICAL & ELECTRONIC
Hoon Hong, Alexey Ovchinnikov, Gleb Pogudin, Chee Yap
{"title":"“微分模型的全局可辨识性”的勘误","authors":"Hoon Hong,&nbsp;Alexey Ovchinnikov,&nbsp;Gleb Pogudin,&nbsp;Chee Yap","doi":"10.1002/cpa.22163","DOIUrl":null,"url":null,"abstract":"<p>We are grateful to Peter Thompson for pointing out an error in [<span>1</span>, Lemma 3.5, p. 1848]. The original proof worked only under the assumption that <math>\n <semantics>\n <mover>\n <mi>θ</mi>\n <mo>̂</mo>\n </mover>\n <annotation>$\\hat{\\theta }$</annotation>\n </semantics></math> is a vector of constants. However, some of the components of <math>\n <semantics>\n <mover>\n <mi>θ</mi>\n <mo>̂</mo>\n </mover>\n <annotation>$\\hat{\\bm{\\theta }}$</annotation>\n </semantics></math> could be the states of the dynamic under consideration, and the lemma was used in such a setup (i.e., with <math>\n <semantics>\n <mover>\n <mi>θ</mi>\n <mo>̂</mo>\n </mover>\n <annotation>$\\hat{\\bm{\\theta }}$</annotation>\n </semantics></math> involving states) later in [<span>1</span>, Proposition 3.4].</p><p>We give a more explicit version of the statement and provide a correct proof. The desired statement will be deduced from the following:\n\n </p><p>\n </p><p>The following corollary is equivalent to [<span>1</span>, Lemma 3.5, p. 1848] but explicitly highlights that some of the entries of <math>\n <semantics>\n <mover>\n <mi>θ</mi>\n <mo>̂</mo>\n </mover>\n <annotation>$\\hat{\\bm{\\theta }}$</annotation>\n </semantics></math> may be initial conditions, not only system parameters.\n\n </p><p>\n </p>","PeriodicalId":3,"journal":{"name":"ACS Applied Electronic Materials","volume":null,"pages":null},"PeriodicalIF":4.3000,"publicationDate":"2023-09-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://onlinelibrary.wiley.com/doi/epdf/10.1002/cpa.22163","citationCount":"0","resultStr":"{\"title\":\"Erratum for “Global Identifiability of Differential Models”\",\"authors\":\"Hoon Hong,&nbsp;Alexey Ovchinnikov,&nbsp;Gleb Pogudin,&nbsp;Chee Yap\",\"doi\":\"10.1002/cpa.22163\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"<p>We are grateful to Peter Thompson for pointing out an error in [<span>1</span>, Lemma 3.5, p. 1848]. The original proof worked only under the assumption that <math>\\n <semantics>\\n <mover>\\n <mi>θ</mi>\\n <mo>̂</mo>\\n </mover>\\n <annotation>$\\\\hat{\\\\theta }$</annotation>\\n </semantics></math> is a vector of constants. However, some of the components of <math>\\n <semantics>\\n <mover>\\n <mi>θ</mi>\\n <mo>̂</mo>\\n </mover>\\n <annotation>$\\\\hat{\\\\bm{\\\\theta }}$</annotation>\\n </semantics></math> could be the states of the dynamic under consideration, and the lemma was used in such a setup (i.e., with <math>\\n <semantics>\\n <mover>\\n <mi>θ</mi>\\n <mo>̂</mo>\\n </mover>\\n <annotation>$\\\\hat{\\\\bm{\\\\theta }}$</annotation>\\n </semantics></math> involving states) later in [<span>1</span>, Proposition 3.4].</p><p>We give a more explicit version of the statement and provide a correct proof. The desired statement will be deduced from the following:\\n\\n </p><p>\\n </p><p>The following corollary is equivalent to [<span>1</span>, Lemma 3.5, p. 1848] but explicitly highlights that some of the entries of <math>\\n <semantics>\\n <mover>\\n <mi>θ</mi>\\n <mo>̂</mo>\\n </mover>\\n <annotation>$\\\\hat{\\\\bm{\\\\theta }}$</annotation>\\n </semantics></math> may be initial conditions, not only system parameters.\\n\\n </p><p>\\n </p>\",\"PeriodicalId\":3,\"journal\":{\"name\":\"ACS Applied Electronic Materials\",\"volume\":null,\"pages\":null},\"PeriodicalIF\":4.3000,\"publicationDate\":\"2023-09-22\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"https://onlinelibrary.wiley.com/doi/epdf/10.1002/cpa.22163\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"ACS Applied Electronic Materials\",\"FirstCategoryId\":\"100\",\"ListUrlMain\":\"https://onlinelibrary.wiley.com/doi/10.1002/cpa.22163\",\"RegionNum\":3,\"RegionCategory\":\"材料科学\",\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"Q1\",\"JCRName\":\"ENGINEERING, ELECTRICAL & ELECTRONIC\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"ACS Applied Electronic Materials","FirstCategoryId":"100","ListUrlMain":"https://onlinelibrary.wiley.com/doi/10.1002/cpa.22163","RegionNum":3,"RegionCategory":"材料科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"Q1","JCRName":"ENGINEERING, ELECTRICAL & ELECTRONIC","Score":null,"Total":0}
引用次数: 0

摘要

我们感谢彼得·汤普森指出了[1,引理3.5,第1848页]中的一个错误。最初的证明只有在θ´$\hat{\ θ}$是一个常数向量的假设下才有效。然而,θ θ $\hat{\bm{\theta}}$的一些分量可以是所考虑的动态状态,引理在后面的[1,命题3.4]中被用于这样的设置(即θ θ $\hat{\bm{\theta}}$涉及状态)。我们给出了一个更明确的版本,并提供了一个正确的证明。我们想要的表述将从下面推导出来:引理1。考虑一个微分方程组x1 ' = f1 (x,μ,u),⋮xn”= fn (x,μ,u), $ $ \开始{方程}{\{病例}开始x_1 ^ {\ '} = f (\ bm {x} \ bm{\μ}\ bm{你}),\ \ \ vdots \ \ x_n ^ {\ '} = fn (\ bm {x} \ bm{\μ}\ bm{你}),结束\{病例}}\{方程}$ $结束(1)x = (x1,…,xn)美元\ bm {x} = (x_1、\ ldots x_n)美元和u = (u1,……,嗯)美元\ bm{你}= (u_1 \ ldots, u_m)美元的元组微分不定,μ=(μ1,…,μλ)美元\ bm{\μ}=(\μ_1 \ ldots \μ_ \λ)美元是标量参数,和f1,…,fn∈C (x,μ,u) $ f \ ldots,f_n \in \mathbb {C}(\bm{x}, \bm{\mu}, \bm{u})$。让问(x,μ,u)∈C (x,μ,u)问美元(\ bm {x} \ bm{\μ}\ bm{你})中\ \ mathbb {C} [\ bm {x} \ bm{\μ}\ bm{你}]美元的LCM f1的分母,…,fn $ f \ ldots fn美元。设P∈C[x,μ]{u}$P \in \mathbb {C}[\bm{x}, \bm{\mu}]\lbrace \bm{u}\rbrace$是一个非零微分多项式。然后存在非零P1∈C (x,μ)P_1美元\ \ mathbb {C} [\ * bm {x} \ bm{\μ}]美元和P2∈C{你}$ P_2 \ \ mathbb {C} \ lbrace \ bm{你}\ rbrace美元,每一个元组μ̂Cλ∈\美元的帽子{\ bm{\μ}}\中\ mathbb {C} ^ \λ美元每个幂级数解(x̂,u) $ ({\ * bm {x}} \帽子,帽子\ {\ bm{你}})美元(1)的参数μ̂美元\帽子{\ bm{\μ}}在C [[t]]美元\ mathbb {C} [\ [t] \ !]美元这样问(x̂,μ̂,u) | t = 0≠0 $ $ \{方程*}开始问(\帽子{\ * bm {x}} \帽子{\ bm{\μ}},\帽子{\ bm{你}})| _ {t = 0} \ 0 ne \{方程*}$ $结束我们P1 (x̂,μ̂)| t = 0≠0,P2 (u) | t = 0≠0⇒P (x̂,μ̂,u)≠0。$ ${方程*}{\ \开始离开(P_1 ({\ * bm {x}} \帽子,帽子\ {\ bm{\μ}})| _ {t = 0} \ 0 ne \; \, \;P_2(\帽子{\ bm{你}})| _ {t = 0} \ 0 ne \右)}\ Rightarrow P(\帽子{\ * bm {x}} \帽子{\ bm{\μ}},帽子\ {\ bm{你}})\ 0。结束\{方程*}$ $的证明。考虑下面的微分理想我:=⟨(Qxi−Qfi) (j)、P (j)∣1⩽我⩽n, j⩾0⟩:问∞⊂C(μ){x, u}。$ $ \{方程*}我开始:= \ langle (Qx_i ^ {\ '} - Qf_i) ^ {(j)}, P ^ {(j)} \ 1 \ leqslant我\ leqslant n,中期\;j \geqslant 0 \rangle: Q^\ inty \子集\mathbb {C}[\bm{\mu}]\lbrace \bm{x}, \bm{u}\rbrace。我们声明I包含一个形式为P1P2$P_1P_2$的非零多项式,使得P1∈C[x,μ]$P_1 \in \mathbb {C}[\bm{x}, \bm{\mu}]$和P2∈C{u}$P_2 \in \mathbb {C}\lbrace \bm{u}\rbrace$。首先,我们将证明,如果断言为真,那么P1和P2满足引理的条件。相反,假设有一个幂级数解(x̂,u) $ ({\ * bm {x}} \帽子,帽子\ {\ bm{你}})美元(1)的参数μ̂美元\帽子{\ bm{\μ}}$的常数项Q (x̂,μ̂,u) P1 (x̂,μ̂)P2 (u)问美元({\ * bm {x}} \帽子,帽子\ {\ bm{\μ}},帽子\ {\ bm{你}})P_1 ({\ * bm {x}} \帽子,帽子\ {\ bm{\μ}})P_2(\帽子{\ bm{你}})美元零但P (x̂,μ̂,u) = 0 $ P ({\ * bm {x}} \帽子,帽子\ {\ bm{\μ}},帽子\ {\ bm{你}})= 0美元。由于(x³,μ³,û)$(\hat{\bm{x}}, \hat{\bm{\mu}}, \hat{\bm{u}})$是微分多项式P和Qxi ' - Qfi$Qx_i^{\素数}- Qf_i$的零,对于每一个1≤i≤n$1 \leqslant i \leqslant n$,它是理想⟨(Qxi ' - Qfi)(j),P(j)∣1≤i≤n,j≠0⟩的零。$ ${方程*}\ \开始langle (Qx_i ^ {\ '} - Qf_i) ^ {(j)}, P ^ {(j)} \中期1 \ leqslant \ leqslant n \;J \geqslant 0 \rangle。\{方程*}$ $结束以来Q (x̂,μ̂,u) | t = 0≠0美元Q ({\ * bm {x}} \帽子,帽子\ {\ bm{\μ}},帽子\ {\ bm{你}})| _ {t = 0} \ 0美元,在我每一个元素,这是上面的理想的饱和,也消失在(x̂,μ̂,u)美元({\ * bm {x}} \帽子,帽子\ {\ bm{\μ}},帽子\ {\ bm{你}})美元。特别是,P1P2 P_1P_2消失在美元(x̂,μ̂,u)美元(\帽子{\ * bm {x}} \帽子{\ bm{\μ}},\帽子{\ bm{你}})美元,我们到达的矛盾与P1 (x̂,μ̂)P2 (u)≠0美元P_1(\帽子{\ * bm {x}} \帽子{\ bm{\μ}})P_2(\帽子{\ bm{你}})\ 0美元。现在我们来证明这个说法。考虑环R: = C (x,μ){你}[1 / Q] $ R: = \ mathbb {C} [\ * bm {x} \ bm{\μ}]\ lbrace \ bm{你}\ rbrace [1 / Q]美元。设J是r中I∩C[x,μ]{u}$I \cap \mathbb {C}[\bm{x}, \bm{\mu}]\lbrace \bm{u}\rbrace$生成的理想。通过Q处的饱和定义I意味着J∩C[x,μ]{u}=I∩C[x,μ]{u}。$ $ \{方程*}开始J \帽\ mathbb {C} [\ bm {x}, {\ bm{\μ}}]\ lbrace \ bm{你}\ rbrace =我\帽\ mathbb {C} [\ bm {x}, {\ bm{\μ}}]\ lbrace \ bm{你}\ rbrace。因此,足以证明存在一个形式为P1P2$P_1 P_2$的元素,其中P1∈C[x,μ]$P_1 \in \mathbb {C}[\bm{x}, \bm{\mu}]$, P2∈C{u}$P_2 \in \mathbb {C}\lbrace \bm{u}\rbrace$ in j。我们在R上定义一个导数L$\mathcal {L}$(基本上是李氏导数):=∑i=1nfi∂g∂xi+∑j=0∞u R (j+1)∂g∂u R (j)forg∈R。$ ${方程*}\ \开始mathcal {1} (g): = \ \和限制_ {i = 1} ^ n f_i \压裂{\部分g}{\部分x_i} + \ \和限制_{\魔法= 1}^ m \ \和限制_ {j = 0} ^ \ infty u_ \字母L ^ {(j + 1)} \压裂{\部分g}{\部分u_ \魔法^ {(j)}} \四\文本为}{g \ R。 由于Qx1 '−Qf1,…,Qxn '−Qfn∈I$Qx_1^{\素数}- Qf_1, \ldots, Qx_n^{\素数}- Qf_n \in I$和I是微分理想,J在L$\mathcal {L}$下是不变的。设R ~ $\ widdetilde {R}$是R对C{u}$\mathbb {C}\lbrace \bm{u}\rbrace$的定位,J ~ $\ widdetilde {J}$是J在这个定位中产生的理想值。推导L$\mathcal {L}$可以自然地扩展到R ~ $\ widdetilde {R}$,并且J ~ $\ widdetilde {J}$也是L$\mathcal {L}$-不变的。足以证明J ~∩C[x,μ]≠{0}$\ widdetilde {J}\cap \mathbb {C}[\bm{x}, \bm{\mu}] \ne \lbrace 0\rbrace$。考虑一个J ~∩C[x,μ]{u}$\ widdetilde {J} \cap \mathbb {C}[\bm{x}, \bm{\mu}]\lbrace \bm{u}\rbrace$的非零元素,它具有最小的单项式个数,并且在这些元素中,有一个总度最小的元素。我们称之为S。如果S∈C[x,μ]$S\in \mathbb {C}[\bm{x}, \bm{\mu}]$,我们就完成了。否则,u中的一个出现在S中,比如u1。设h=ordu1S$h = \operatorname{ord}_{u_1}S$。由于R ~ $\widetilde{R}$是一个诺瑟环,因此存在N&gt;0$N &gt;0$使得LN(S)∈⟨S,L(S),…,LN−1(S)⟩。$ ${方程*}\ \开始mathcal {L} ^ N (S) \ \ langle年代\ mathcal {1} (S) \ ldots \ mathcal {L} ^ {N - 1} \捕杀。$$我们有ordu1Li(S)&lt;N+h$\operatorname{ord}_{u_1} \mathcal {L}^i(S) &lt;N$i &lt;N$i &lt;N和LN (S) =∂年代美元∂u1 (h) u1 (h + N) + T, whereordu1T&lt; N + h。$ ${方程*}\ \开始mathcal {L} ^ N (S) = \压裂{\部分年代}{\部分u_1 ^ {(h)}} u_1 ^ {(h + N)} + T,{在}\ \四\文本operatorname{奥德}_ {u_1} T & lt;N + h。\{方程*}$ $因此结束,我们有∂年代∂u1 (h)∈⟨年代,L (S),…,LN−1 (S)⟩⊂J∼。$$\begin{equation*} \frac{\partial S}{\partial u_1^{(h)}} \in \langle S, \mathcal {L}(S), \ldots, \mathcal {L}^{N - 1}(S) \rangle \子集\ widdetilde {J}。$$ $如果S可以被u1(h)$u_1^{(h)}$整除,那么S/u1(h)∈J ~ $S / u_1^{(h)} \in \ widdetilde {J}$将具有相同数量的单项式,但阶数更小,这与S的选择相矛盾。因此,∂S∂u1(h)$\frac{\偏S}{\偏u_1^{(h)}}$的单项式比S少,因此与S的选择相矛盾。□$\Box$P. 1848]但明确地强调了θ n $\hat{\bm{\theta}}$的一些条目可能是初始条件,而不仅仅是系统参数。推论1。(澄清版本的[[1],引理3.5,p . 1848])的符号(1,2.2节),让p(μx, u,…,u (N))∈C[μ,x]{你}$ p (\ bm{\μ}\ bm {x}, u, \ ldots u ^ {(N)})中\ \ mathbb {C} [\ * bm{\μ}\ bm {x}] \ lbrace u \ rbrace美元是零。然后存在非空的Zariski开子集Θ⊂Cs美元\θ{\子集}\ mathbb {C} ^{年代}$和U⊂C∞(0)美元U \ \子集mathbb {C} ^ {\ infty}(0),美元每θ̂=(μ̂,x̂*)∈Θ美元\帽子{\ bm{\θ}}=(\帽子{\ bm{\μ}},帽子\ {\ bm {x}} ^ \ ast) \ \θ,美元U∈\帽子{你}\ U的美元,和相应的x̂= x(θ̂,U)美元\帽子{\ bm {x}} = x(\帽子{\ bm{\θ}},{你}\帽子),美元函数P(μ̂,x̂,U,…,(U) (N)) $ P(\帽子{\ bm{\μ}},{\ * bm {x}} \帽子,\帽子{你}\ ldots,帽子(\{你})^ {(N)})是一个非零元素的C∞(0)美元\ mathbb {C} ^ {\ infty} (0) .Proof美元。我们将引理1应用于模型Σ和命题中的多项式P,得到多项式P1(x,μ)$P_1(\bm{x}, \bm{\mu})$和P2(u)。我们分别用P1≠0$P_1 \ne 0$和P2(U)|t=0≠0$P_2(\bm{U})|_{t =0} \ne 0$来定义Zariski开集Θ和U。引理意味着,(μ̂,x̂∗)∈Θ美元(\帽子{\ bm{\μ}},帽子\ {\ bm {x}} ^ *) \中\θ和u∈\美元的帽子在u ${你}\ P(μ̂,x̂,u,…,(u) (N)) $ P(\帽子{\ bm{\μ}},帽子\ {\ bm {x}},{你}\帽子,\ ldots,帽子(\{你})^ {(N)})将一个非零函数。□\美元美元
本文章由计算机程序翻译,如有差异,请以英文原文为准。
Erratum for “Global Identifiability of Differential Models”

We are grateful to Peter Thompson for pointing out an error in [1, Lemma 3.5, p. 1848]. The original proof worked only under the assumption that θ ̂ $\hat{\theta }$ is a vector of constants. However, some of the components of θ ̂ $\hat{\bm{\theta }}$ could be the states of the dynamic under consideration, and the lemma was used in such a setup (i.e., with θ ̂ $\hat{\bm{\theta }}$ involving states) later in [1, Proposition 3.4].

We give a more explicit version of the statement and provide a correct proof. The desired statement will be deduced from the following:

The following corollary is equivalent to [1, Lemma 3.5, p. 1848] but explicitly highlights that some of the entries of θ ̂ $\hat{\bm{\theta }}$ may be initial conditions, not only system parameters.

求助全文
通过发布文献求助,成功后即可免费获取论文全文。 去求助
来源期刊
CiteScore
7.20
自引率
4.30%
发文量
567
×
引用
GB/T 7714-2015
复制
MLA
复制
APA
复制
导出至
BibTeX EndNote RefMan NoteFirst NoteExpress
×
提示
您的信息不完整,为了账户安全,请先补充。
现在去补充
×
提示
您因"违规操作"
具体请查看互助需知
我知道了
×
提示
确定
请完成安全验证×
copy
已复制链接
快去分享给好友吧!
我知道了
右上角分享
点击右上角分享
0
联系我们:info@booksci.cn Book学术提供免费学术资源搜索服务,方便国内外学者检索中英文文献。致力于提供最便捷和优质的服务体验。 Copyright © 2023 布克学术 All rights reserved.
京ICP备2023020795号-1
ghs 京公网安备 11010802042870号
Book学术文献互助
Book学术文献互助群
群 号:481959085
Book学术官方微信