- #1
nweibley
- 5
- 0
Preface: This homework assignment was due long ago. At this point I am only trying to understand the problem (or really if the posted solution follows the problem) before my exam. I have no real indication that this problem (or even one like it) will be on my final, but I feel that my inability to solve it shows a weakness in my understanding of the material that I should fix before I sit down to take the test. The PDF for the assignment is http://csc.list.ufl.edu/3105/fall08/assignment/assignment-8.pdf. So here goes...
This the exact wording of the problem:
[tex]\mathbf{x}[n+1] = A\mathbf{x}[n] + Bu[n][/tex]
[tex]\mathbf{y}[n] = C\mathbf{x}[n] + Du[n][/tex]
Well... first I tried to setup a difference equation of the form:
[tex]\-1y[n+1] + \left(1+\frac{.09}{12}\right) y[n]=u[n][/tex]
and set
[tex]x_{1}[n]=y[n][/tex]
[tex]x_{2}[n]=y[n+1][/tex]
and
[tex]x_{1}[n+1]=y[n+1]=x_{2}[n][/tex]However, I came to realize that this is either a) false or b) going to yield a state space representation with 1 state (A is a 1x1 matrix).
I'd be thrilled with a lead to chase on this one; I can't tell if the question as-asked makes sense insomuch as I'm going to find a system matrix of dimensions m x m (m > 1) which are the only state spaces I can remember working in class.
The posted solutions from the TA show a difference equation of the form:
[tex]x[n+1] = ax[n]-ku[n][/tex]
Which is then Z transformed to yield:
[tex]X(z) = -k\left(\frac{z}{\left(z-1\right)\left(z-a\right)}\right)+\frac{z}{z-a}x[0][/tex]
At which point I believe the TA made a superficial error (but am probably wrong about that).
So, any guidance? Is there a state space solution that makes sense for this problem? And if so, does it essentially present itself like scalars?
Many thanks for any pointers.
Homework Statement
This the exact wording of the problem:
Use a discrete time state space approach to model a mortgage payment
system and find the fixed monthly payment required to amortize
a loan of $54546.00 at 9% per annum over a period of 60 months.
(Answer should be 1132.29 using effective monthly interest as .09/12)
Homework Equations
[tex]\mathbf{x}[n+1] = A\mathbf{x}[n] + Bu[n][/tex]
[tex]\mathbf{y}[n] = C\mathbf{x}[n] + Du[n][/tex]
The Attempt at a Solution
Well... first I tried to setup a difference equation of the form:
[tex]\-1y[n+1] + \left(1+\frac{.09}{12}\right) y[n]=u[n][/tex]
and set
[tex]x_{1}[n]=y[n][/tex]
[tex]x_{2}[n]=y[n+1][/tex]
and
[tex]x_{1}[n+1]=y[n+1]=x_{2}[n][/tex]However, I came to realize that this is either a) false or b) going to yield a state space representation with 1 state (A is a 1x1 matrix).
I'd be thrilled with a lead to chase on this one; I can't tell if the question as-asked makes sense insomuch as I'm going to find a system matrix of dimensions m x m (m > 1) which are the only state spaces I can remember working in class.
The posted solutions from the TA show a difference equation of the form:
[tex]x[n+1] = ax[n]-ku[n][/tex]
Which is then Z transformed to yield:
[tex]X(z) = -k\left(\frac{z}{\left(z-1\right)\left(z-a\right)}\right)+\frac{z}{z-a}x[0][/tex]
At which point I believe the TA made a superficial error (but am probably wrong about that).
So, any guidance? Is there a state space solution that makes sense for this problem? And if so, does it essentially present itself like scalars?
Many thanks for any pointers.
Last edited by a moderator: