Characterizing the frontier between upward sloping and downward sloping parts of a curve
up vote
1
down vote
favorite
Let me start with an example, let
$$mathbf{T}=begin{bmatrix}2&1\1&3end{bmatrix}$$
and $α=(2,2)$. Abusing notation, write $mathbf{λ}^{α}=(λ^2,(1-λ)^2)$ for any $0≤λ≤1$. If we compute $T^{top}mathbf{λ}^{α}$ for every value of $λ$, we obtain a series of points $(x,y)$ that connect $(1,2)$ and $(3,1)$ through a curve. As long as for all $i$, $α_i>1,$ this curve goes through a point in which $x$ is minimum (call it $minX$) and another in which $y$ is minimum ($minY$). In the example, $minX=(frac{3}{4},frac{19}{16})$ and $minY=(frac{13}{9},frac{2}{3}),$ that correspond to weights $(frac{3}{4},frac{1}{4})$ and $(frac{1}{3},frac{2}{3}),$ respectively.
I'm only interested in those $mathbf{λ}$ that yield points between $minX$ and $minY,$ i.e., points that are the result of $frac{1}{3}leqλleqfrac{3}{4}$. In these points, the slope of the curve is negative and $minX$ and $minY$ are the "frontier" points that separate the part in which the curve is negatively sloped from those where it is not.
Similarly, for $n>2$, I'd be interested in those points sitting in the part of curved surface that is "negatively sloped". In $Bbb R^3$, I want to exclude those portions of the surface for which $x,y$ and $z$ increase simultaneously. Of course, I can still find values for $minX,$ $minY$ and $minZ,$ but I don't know how to determine the rest of the "frontier" points.
Additional information: If we define $mathbf{k}=mathbf{T}^{top}mathbf{barλ}^{α}$ for some $mathbf{barλ}=(barλ,(1-barλ))$, and $0leq barλ leq 1$, the $mathbf{λ}$ yielding points between $minX$ and $minY$ have an additional property: they are the ones for which the solution to the problem
$$max sum_i lambda_i$$
$$s.t. T^{top}λ^{α}=mathbf{k},$$
denoted $mathbf{λ}^*$, has the property that $sum_i lambda_i^*=1.$ Observe that $mathbf{barλ}$ is part of the feasible set, and therefore $sum_i lambda_i^*geq1.$ What I would like is to characterize the $mathbf{λ}$ that are in the frontier between those resulting in $sum_i lambda_i^*=1$ and in $sum_i lambda_i^*>1.$
(I'm not sure about the tag, so I apologize if I've misled you and I'd be happy to change it if you can suggest a better one).
algebraic-geometry
asked Nov 6 at 14:09
Patricio
1165
|
show 4 more comments
up vote
1
down vote
favorite
Let me start with an example, let
$$mathbf{T}=begin{bmatrix}2&1\1&3end{bmatrix}$$
and $α=(2,2)$. Abusing notation, write $mathbf{λ}^{α}=(λ^2,(1-λ)^2)$ for any $0≤λ≤1$. If we compute $T^{top}mathbf{λ}^{α}$ for every value of $λ$, we obtain a series of points $(x,y)$ that connect $(1,2)$ and $(3,1)$ through a curve. As long as for all $i$, $α_i>1,$ this curve goes through a point in which $x$ is minimum (call it $minX$) and another in which $y$ is minimum ($minY$). In the example, $minX=(frac{3}{4},frac{19}{16})$ and $minY=(frac{13}{9},frac{2}{3}),$ that correspond to weights $(frac{3}{4},frac{1}{4})$ and $(frac{1}{3},frac{2}{3}),$ respectively.
I'm only interested in those $mathbf{λ}$ that yield points between $minX$ and $minY,$ i.e., points that are the result of $frac{1}{3}leqλleqfrac{3}{4}$. In these points, the slope of the curve is negative and $minX$ and $minY$ are the "frontier" points that separate the part in which the curve is negatively sloped from those where it is not.
Similarly, for $n>2$, I'd be interested in those points sitting in the part of curved surface that is "negatively sloped". In $Bbb R^3$, I want to exclude those portions of the surface for which $x,y$ and $z$ increase simultaneously. Of course, I can still find values for $minX,$ $minY$ and $minZ,$ but I don't know how to determine the rest of the "frontier" points.
Additional information: If we define $mathbf{k}=mathbf{T}^{top}mathbf{barλ}^{α}$ for some $mathbf{barλ}=(barλ,(1-barλ))$, and $0leq barλ leq 1$, the $mathbf{λ}$ yielding points between $minX$ and $minY$ have an additional property: they are the ones for which the solution to the problem
$$max sum_i lambda_i$$
$$s.t. T^{top}λ^{α}=mathbf{k},$$
denoted $mathbf{λ}^*$, has the property that $sum_i lambda_i^*=1.$ Observe that $mathbf{barλ}$ is part of the feasible set, and therefore $sum_i lambda_i^*geq1.$ What I would like is to characterize the $mathbf{λ}$ that are in the frontier between those resulting in $sum_i lambda_i^*=1$ and in $sum_i lambda_i^*>1.$
(I'm not sure about the tag, so I apologize if I've misled you and I'd be happy to change it if you can suggest a better one).
algebraic-geometry
asked Nov 6 at 14:09
Patricio
1165
There are some things that need to be cleaned up I think. If $alpha$ is an ordered pair, then what do you mean by "As long as $alpha > 1$? Why are you taking the transpose of the symmetric matrix $T$? And lastly, it seems to me that for the example given, the curve is a straight line segment between $(1,3)$ and $(2,1)$.
– dbx
Nov 6 at 14:33
@dbx, I meant that each component of $α$ needed to be larger than $1$. I've edited my question and hope it is clearer now. I need to use the transpose because I want to combine columns rather than rows, is just an unhappy coincidence that I've chosen a symmetric matrix. I've corrected also the definition of $λ^{α},$ so it is now the curve it was meant to be. Thank you very much.
– Patricio
Nov 6 at 15:10
I've been playing around with this very interesting problem, and I don't have a solution -- but I will note that Mathematica also doesn't know how to solve it in general.
– dbx
Nov 6 at 16:10
@dbx, how did you ask Mathematica?
– Patricio
Nov 7 at 7:46
I didn't save my code, but it's straightforward to write $Tcdot lambda^alpha$ as a system of two nonlinear equations in $lambda$. Their derivatives are easy enough to find, and finding your frontier points corresponds to finding their critical points. I only tried the Solve[ ] function, which failed.
– dbx
Nov 7 at 18:45
|
show 4 more comments
up vote
1
down vote
favorite
up vote
1
down vote
favorite
Let me start with an example, let
$$mathbf{T}=begin{bmatrix}2&1\1&3end{bmatrix}$$
and $α=(2,2)$. Abusing notation, write $mathbf{λ}^{α}=(λ^2,(1-λ)^2)$ for any $0≤λ≤1$. If we compute $T^{top}mathbf{λ}^{α}$ for every value of $λ$, we obtain a series of points $(x,y)$ that connect $(1,2)$ and $(3,1)$ through a curve. As long as for all $i$, $α_i>1,$ this curve goes through a point in which $x$ is minimum (call it $minX$) and another in which $y$ is minimum ($minY$). In the example, $minX=(frac{3}{4},frac{19}{16})$ and $minY=(frac{13}{9},frac{2}{3}),$ that correspond to weights $(frac{3}{4},frac{1}{4})$ and $(frac{1}{3},frac{2}{3}),$ respectively.
I'm only interested in those $mathbf{λ}$ that yield points between $minX$ and $minY,$ i.e., points that are the result of $frac{1}{3}leqλleqfrac{3}{4}$. In these points, the slope of the curve is negative and $minX$ and $minY$ are the "frontier" points that separate the part in which the curve is negatively sloped from those where it is not.
Similarly, for $n>2$, I'd be interested in those points sitting in the part of curved surface that is "negatively sloped". In $Bbb R^3$, I want to exclude those portions of the surface for which $x,y$ and $z$ increase simultaneously. Of course, I can still find values for $minX,$ $minY$ and $minZ,$ but I don't know how to determine the rest of the "frontier" points.
Additional information: If we define $mathbf{k}=mathbf{T}^{top}mathbf{barλ}^{α}$ for some $mathbf{barλ}=(barλ,(1-barλ))$, and $0leq barλ leq 1$, the $mathbf{λ}$ yielding points between $minX$ and $minY$ have an additional property: they are the ones for which the solution to the problem
$$max sum_i lambda_i$$
$$s.t. T^{top}λ^{α}=mathbf{k},$$
denoted $mathbf{λ}^*$, has the property that $sum_i lambda_i^*=1.$ Observe that $mathbf{barλ}$ is part of the feasible set, and therefore $sum_i lambda_i^*geq1.$ What I would like is to characterize the $mathbf{λ}$ that are in the frontier between those resulting in $sum_i lambda_i^*=1$ and in $sum_i lambda_i^*>1.$
(I'm not sure about the tag, so I apologize if I've misled you and I'd be happy to change it if you can suggest a better one).
algebraic-geometry
asked Nov 6 at 14:09
Patricio
1165
Let me start with an example, let
$$mathbf{T}=begin{bmatrix}2&1\1&3end{bmatrix}$$
and $α=(2,2)$. Abusing notation, write $mathbf{λ}^{α}=(λ^2,(1-λ)^2)$ for any $0≤λ≤1$. If we compute $T^{top}mathbf{λ}^{α}$ for every value of $λ$, we obtain a series of points $(x,y)$ that connect $(1,2)$ and $(3,1)$ through a curve. As long as for all $i$, $α_i>1,$ this curve goes through a point in which $x$ is minimum (call it $minX$) and another in which $y$ is minimum ($minY$). In the example, $minX=(frac{3}{4},frac{19}{16})$ and $minY=(frac{13}{9},frac{2}{3}),$ that correspond to weights $(frac{3}{4},frac{1}{4})$ and $(frac{1}{3},frac{2}{3}),$ respectively.
I'm only interested in those $mathbf{λ}$ that yield points between $minX$ and $minY,$ i.e., points that are the result of $frac{1}{3}leqλleqfrac{3}{4}$. In these points, the slope of the curve is negative and $minX$ and $minY$ are the "frontier" points that separate the part in which the curve is negatively sloped from those where it is not.
Similarly, for $n>2$, I'd be interested in those points sitting in the part of curved surface that is "negatively sloped". In $Bbb R^3$, I want to exclude those portions of the surface for which $x,y$ and $z$ increase simultaneously. Of course, I can still find values for $minX,$ $minY$ and $minZ,$ but I don't know how to determine the rest of the "frontier" points.
Additional information: If we define $mathbf{k}=mathbf{T}^{top}mathbf{barλ}^{α}$ for some $mathbf{barλ}=(barλ,(1-barλ))$, and $0leq barλ leq 1$, the $mathbf{λ}$ yielding points between $minX$ and $minY$ have an additional property: they are the ones for which the solution to the problem
$$max sum_i lambda_i$$
$$s.t. T^{top}λ^{α}=mathbf{k},$$
denoted $mathbf{λ}^*$, has the property that $sum_i lambda_i^*=1.$ Observe that $mathbf{barλ}$ is part of the feasible set, and therefore $sum_i lambda_i^*geq1.$ What I would like is to characterize the $mathbf{λ}$ that are in the frontier between those resulting in $sum_i lambda_i^*=1$ and in $sum_i lambda_i^*>1.$
(I'm not sure about the tag, so I apologize if I've misled you and I'd be happy to change it if you can suggest a better one).
algebraic-geometry
algebraic-geometry
asked Nov 6 at 14:09
Patricio
1165
asked Nov 6 at 14:09
Patricio
1165
edited Nov 26 at 12:32
asked Nov 6 at 14:09
Patricio
1165
asked Nov 6 at 14:09
Patricio
1165
asked Nov 6 at 14:09
Patricio
1165
1165
There are some things that need to be cleaned up I think. If $alpha$ is an ordered pair, then what do you mean by "As long as $alpha > 1$? Why are you taking the transpose of the symmetric matrix $T$? And lastly, it seems to me that for the example given, the curve is a straight line segment between $(1,3)$ and $(2,1)$.
– dbx
Nov 6 at 14:33
@dbx, I meant that each component of $α$ needed to be larger than $1$. I've edited my question and hope it is clearer now. I need to use the transpose because I want to combine columns rather than rows, is just an unhappy coincidence that I've chosen a symmetric matrix. I've corrected also the definition of $λ^{α},$ so it is now the curve it was meant to be. Thank you very much.
– Patricio
Nov 6 at 15:10
I've been playing around with this very interesting problem, and I don't have a solution -- but I will note that Mathematica also doesn't know how to solve it in general.
– dbx
Nov 6 at 16:10
@dbx, how did you ask Mathematica?
– Patricio
Nov 7 at 7:46
I didn't save my code, but it's straightforward to write $Tcdot lambda^alpha$ as a system of two nonlinear equations in $lambda$. Their derivatives are easy enough to find, and finding your frontier points corresponds to finding their critical points. I only tried the Solve[ ] function, which failed.
– dbx
Nov 7 at 18:45
|
show 4 more comments
There are some things that need to be cleaned up I think. If $alpha$ is an ordered pair, then what do you mean by "As long as $alpha > 1$? Why are you taking the transpose of the symmetric matrix $T$? And lastly, it seems to me that for the example given, the curve is a straight line segment between $(1,3)$ and $(2,1)$.
– dbx
Nov 6 at 14:33
@dbx, I meant that each component of $α$ needed to be larger than $1$. I've edited my question and hope it is clearer now. I need to use the transpose because I want to combine columns rather than rows, is just an unhappy coincidence that I've chosen a symmetric matrix. I've corrected also the definition of $λ^{α},$ so it is now the curve it was meant to be. Thank you very much.
– Patricio
Nov 6 at 15:10
I've been playing around with this very interesting problem, and I don't have a solution -- but I will note that Mathematica also doesn't know how to solve it in general.
– dbx
Nov 6 at 16:10
@dbx, how did you ask Mathematica?
– Patricio
Nov 7 at 7:46
I didn't save my code, but it's straightforward to write $Tcdot lambda^alpha$ as a system of two nonlinear equations in $lambda$. Their derivatives are easy enough to find, and finding your frontier points corresponds to finding their critical points. I only tried the Solve[ ] function, which failed.
– dbx
Nov 7 at 18:45
There are some things that need to be cleaned up I think. If $alpha$ is an ordered pair, then what do you mean by "As long as $alpha > 1$? Why are you taking the transpose of the symmetric matrix $T$? And lastly, it seems to me that for the example given, the curve is a straight line segment between $(1,3)$ and $(2,1)$.
– dbx
Nov 6 at 14:33
There are some things that need to be cleaned up I think. If $alpha$ is an ordered pair, then what do you mean by "As long as $alpha > 1$? Why are you taking the transpose of the symmetric matrix $T$? And lastly, it seems to me that for the example given, the curve is a straight line segment between $(1,3)$ and $(2,1)$.
– dbx
Nov 6 at 14:33
@dbx, I meant that each component of $α$ needed to be larger than $1$. I've edited my question and hope it is clearer now. I need to use the transpose because I want to combine columns rather than rows, is just an unhappy coincidence that I've chosen a symmetric matrix. I've corrected also the definition of $λ^{α},$ so it is now the curve it was meant to be. Thank you very much.
– Patricio
Nov 6 at 15:10
@dbx, I meant that each component of $α$ needed to be larger than $1$. I've edited my question and hope it is clearer now. I need to use the transpose because I want to combine columns rather than rows, is just an unhappy coincidence that I've chosen a symmetric matrix. I've corrected also the definition of $λ^{α},$ so it is now the curve it was meant to be. Thank you very much.
– Patricio
Nov 6 at 15:10
I've been playing around with this very interesting problem, and I don't have a solution -- but I will note that Mathematica also doesn't know how to solve it in general.
– dbx
Nov 6 at 16:10
I've been playing around with this very interesting problem, and I don't have a solution -- but I will note that Mathematica also doesn't know how to solve it in general.
– dbx
Nov 6 at 16:10
@dbx, how did you ask Mathematica?
– Patricio
Nov 7 at 7:46
@dbx, how did you ask Mathematica?
– Patricio
Nov 7 at 7:46
I didn't save my code, but it's straightforward to write $Tcdot lambda^alpha$ as a system of two nonlinear equations in $lambda$. Their derivatives are easy enough to find, and finding your frontier points corresponds to finding their critical points. I only tried the Solve[ ] function, which failed.
– dbx
Nov 7 at 18:45
I didn't save my code, but it's straightforward to write $Tcdot lambda^alpha$ as a system of two nonlinear equations in $lambda$. Their derivatives are easy enough to find, and finding your frontier points corresponds to finding their critical points. I only tried the Solve[ ] function, which failed.
– dbx
Nov 7 at 18:45
|
show 4 more comments
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
active
oldest
votes
Thanks for contributing an answer to Mathematics Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Some of your past answers have not been well-received, and you're in danger of being blocked from answering.
Please pay close attention to the following guidance:
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2987216%2fcharacterizing-the-frontier-between-upward-sloping-and-downward-sloping-parts-of%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
There are some things that need to be cleaned up I think. If $alpha$ is an ordered pair, then what do you mean by "As long as $alpha > 1$? Why are you taking the transpose of the symmetric matrix $T$? And lastly, it seems to me that for the example given, the curve is a straight line segment between $(1,3)$ and $(2,1)$.
– dbx
Nov 6 at 14:33
@dbx, I meant that each component of $α$ needed to be larger than $1$. I've edited my question and hope it is clearer now. I need to use the transpose because I want to combine columns rather than rows, is just an unhappy coincidence that I've chosen a symmetric matrix. I've corrected also the definition of $λ^{α},$ so it is now the curve it was meant to be. Thank you very much.
– Patricio
Nov 6 at 15:10
I've been playing around with this very interesting problem, and I don't have a solution -- but I will note that Mathematica also doesn't know how to solve it in general.
– dbx
Nov 6 at 16:10
@dbx, how did you ask Mathematica?
– Patricio
Nov 7 at 7:46
I didn't save my code, but it's straightforward to write $Tcdot lambda^alpha$ as a system of two nonlinear equations in $lambda$. Their derivatives are easy enough to find, and finding your frontier points corresponds to finding their critical points. I only tried the Solve[ ] function, which failed.
– dbx
Nov 7 at 18:45