ScalingStacks

2.2. Implicit function theorem [05DE]

Original official author HTML, exact retained edition. Historical TeX conversion verdicts remain unchanged. Cited-edition alignment and mathematical self-containment are not assessed.

Complete original source context ยท Original author HTML

2.2. Implicit function theorem

For studying the deformation of special lagrangian fibrations, we need the following quantity version of implicit function theorem.

Theorem 2.3 (Theorem 3.2 in [31]).

Let (๐”…1,โˆฅโ‹…โˆฅ1)(\mathfrak{B}_{1},\|\cdot\|_{1}) and (๐”…2,โˆฅโ‹…โˆฅ2)(\mathfrak{B}_{2},\|\cdot\|_{2}) be two Banach spaces, โˆฅโ‹…โˆฅE\|\cdot\|_{E} be the standard Euclidean metric on โ„n\mathbb{R}^{n}, UโŠ‚โ„nร—๐”…1U\subset\mathbb{R}^{n}\times\mathfrak{B}_{1} be an open set, and ๐”‰:UโŸถ๐”…2\mathfrak{F}:U\longrightarrow\mathfrak{B}_{2} be a continuously differentiable map. Denote the differential

Dโ€‹๐”‰โ€‹(y,ฯƒ)โ€‹(yห™+ฯƒห™)=Dyโ€‹๐”‰โ€‹(y,ฯƒ)โ€‹yห™+Dฯƒโ€‹๐”‰โ€‹(y,ฯƒ)โ€‹ฯƒห™,D\mathfrak{F}(y,\sigma)(\dot{y}+\dot{\sigma})=D_{y}\mathfrak{F}(y,\sigma)\dot{y}+D_{\sigma}\mathfrak{F}(y,\sigma)\dot{\sigma},

for (y,ฯƒ)โˆˆU(y,\sigma)\in U, yห™โˆˆโ„n\dot{y}\in\mathbb{R}^{n} and ฯƒห™โˆˆ๐”…1\dot{\sigma}\in\mathfrak{B}_{1}. Assume that (0,0)โˆˆU(0,0)\in U satisfies that Dฯƒโ€‹๐”‰โ€‹(0,0):๐”…1โŸถ๐”…2D_{\sigma}\mathfrak{F}(0,0):\mathfrak{B}_{1}\longrightarrow\mathfrak{B}_{2} has a bounded linear inverse Dฯƒโ€‹๐”‰โ€‹(0,0)โˆ’1:๐”…2โŸถ๐”…1D_{\sigma}\mathfrak{F}(0,0)^{-1}:\mathfrak{B}_{2}\longrightarrow\mathfrak{B}_{1} with

โ€–Dฯƒโ€‹๐”‰โ€‹(0,0)โˆ’1โ€–โ‰คCยฏ\|D_{\sigma}\mathfrak{F}(0,0)^{-1}\|\leq\overline{C}

for a constant Cยฏ>0\overline{C}>0. Let r>0r>0, ฮด0>ฮด>0\delta_{0}>\delta>0 be constants such that, if โ€–y0โ€–E<r,\|y_{0}\|_{E}<r, and โ€–ฯƒโ€–1โ‰คฮด0\|\sigma\|_{1}\leq\delta_{0}, then (y0,ฯƒ)โˆˆU(y_{0},\sigma)\in U, and

โ€–Dฯƒโ€‹๐”‰โ€‹(y0,ฯƒ)โˆ’Dฯƒโ€‹๐”‰โ€‹(0,0)โ€–โ‰ค12โ€‹Cยฏandโ€–๐”‰โก(y0,0)โ€–2โ‰คฮด4โ€‹Cยฏ.\|D_{\sigma}\mathfrak{F}(y_{0},\sigma)-D_{\sigma}\mathfrak{F}(0,0)\|\leq\frac{1}{2\overline{C}}\ \ {\rm and}\ \ \|\mathfrak{F}(y_{0},0)\|_{2}\leq\frac{\delta}{4\overline{C}}.

Then, for any โ€–yโ€–E<r\|y\|_{E}<r, there exists a unique ฯƒโก(y)โˆˆ๐”…1\sigma(y)\in\mathfrak{B}_{1} such that

๐”‰โก(y,ฯƒโก(y))=0,โ€–ฯƒโก(y)โ€–1โ‰คฮด.\mathfrak{F}(y,\sigma(y))=0,\ \ \ \|\sigma(y)\|_{1}\leq\delta.

Furthermore,

Dโ€‹ฯƒโ€‹(y)โ€‹yห™=โˆ’Dฯƒโ€‹๐”‰โ€‹(y,ฯƒ)โˆ’1โ€‹Dyโ€‹๐”‰โ€‹(y,ฯƒ)โ€‹yห™.D\sigma(y)\dot{y}=-D_{\sigma}\mathfrak{F}(y,\sigma)^{-1}D_{y}\mathfrak{F}(y,\sigma)\dot{y}.

The difference between this version of implicit function theorem and the usual one (c.f. [20]) is that we use the condition โ€–๐”‰โก(y,0)โ€–2โ‰คฮด4โ€‹Cยฏ\|\mathfrak{F}(y,0)\|_{2}\leq\frac{\delta}{4\overline{C}} to replace the condition ๐”‰โก(0,0)=0\mathfrak{F}(0,0)=0 besides other quantity estimates.

Original mathematics by the credited authors. Source-backed reader collection; mathematical self-containment is not assessed.