This blog is only for my personal review of mathematical modeling module
1.Scenario
Specifications:
Study4Less, a Student Loans Company, can provide the maintenance aspect of student finance (e.g. food, transport, school materials, utilities, etc). The company loans students a given sum at the beginning of each of the three or four academic years covering the duration of the degree.
Repayment of the loan will start once the graduate is working and earning above a certain (cut-off) salary. Repayments are collected through National Insurance Contributions in the UK and the amount repaid by an individual depends on marginal income earned above the cut-off.
Study4Less loans include a 0.25% monthly interest over the compounded sum borrowed.
Problems:
-
Create a mathematical model to calculate, given a number of months, how much money a student will have to pay monthy.
-
Use your model to calculate how much a student who was in a 4-year programme and borrowed £2000 per annum from Study4Less will have to pay if they want to settle their debt in 10 years and interest .
2.Detailed steps
The problem is supposed to be resolved in the order of steps giving in the lecture.
The steps are:
- problem Identification
- system analysis
- Formulate a mathematical model
- solve the mathmatical model
- Interpret The Model
first of all, what is the system in this scenario? Firstly, we are supposed to figure out what are involved in our system. This could help us set up parameters much more easier.
The system could involved the following:
- M: the amount of money study4less lend to student
- r: monthly interest.
- T: total time allocted for student to settle the loan
- P: monthly payment.
Note that although the monthly interest is fixed, but the monthly payment is decreasing as installments are completed. Therefore, how much a student should pay at month N is depend on how much him or her paid previously. And this is relationship at least we can acquire from the specification.
This question is obviously based on time,just like mentioned above, because the output at time N is based on the output at time N-1. It is always helpful to manually write down the expressions for the first severl time nodes, in this case, we tend to write down, the expression of how much a student left to pay at Nth month.
Assuming A i A_i Ai is the amount of money a student needs to pay after finishing the ith installment. Therefore:
-
A1 (After the first installment)
A 1 = M + M ⋅ r − P A_1 =M+M\cdot r-P A1=M+M⋅r−P
meaning after the first installment a student remain the amount of A1 to pay. -
A2 (similiarly, After the second installment)
A 2 = A 1 + A 1 ⋅ r − P A_2=A_1+A_1 \cdot r-P A2=A1+A1⋅r−P
And we keep doing this for one more month with the hope of finding a pattern that can be mathematically expressed, or in another word, finding a general formula that covers the targeted one.
- A3
A 3 = A 2 + A 2 ⋅ r − P A_3=A_2+A_2 \cdot r-P A3=A2+A2⋅r−P
Observing the expressions above and we can easily notice that the amount of money left to pay at time N is based on that at time N-1. What would happe if we replace A n − 1 A_{n-1} An−1 in the expression at time N with the acutal result?
Take
A
2
A_2
A2 as an example, we replace
A
1
A1
A1 with its expression, we get the following
A
2
=
(
M
+
M
⋅
r
−
P
)
(
1
+
r
)
−
P
A2 = (M+M\cdot r-P)(1+r)-P
A2=(M+M⋅r−P)(1+r)−P
we do the same for A3:
A
3
=
[
(
M
+
M
⋅
r
−
P
)
(
1
+
r
)
−
P
]
(
1
−
r
)
−
P
=
[
(
M
(
1
+
r
)
−
P
)
(
1
+
r
)
−
P
]
(
1
−
r
)
−
P
A3 =[ (M+M\cdot r-P)(1+r)-P](1-r)-P =[ (M(1+r)-P)(1+r)-P](1-r)-P
A3=[(M+M⋅r−P)(1+r)−P](1−r)−P=[(M(1+r)−P)(1+r)−P](1−r)−P
we can do these steps as many as we want to ensure the pattern we found is really capable of decribing the problem. However,for this case, three attempts are enough to tell the pattern.
Noticeably, there are increasingly number of repetitive factors showing in the expressions. And this is where simplification can be applied.
Then we extend the expression of A3 by multipying every (1-r) in to brackets. Then we have the following:
A3 =
M
(
1
+
r
)
3
−
P
(
1
+
r
)
2
−
P
(
1
+
r
)
−
P
M(1+r)^3-P(1+r)^2-P(1+r)-P
M(1+r)3−P(1+r)2−P(1+r)−P
This is the expression of A3 after a litte bit of transformation. If we continue to calculate the expression of A4,A5, you will found that they have more terms and high order, but still follows the pattern. Therefore, we can come into a conculsion and formulate a general expression to cover all possibilities in this scenario:
A
n
=
M
(
1
+
r
)
n
−
P
(
1
+
r
)
n
−
1
−
P
(
1
+
r
)
n
−
2
−
.
.
.
.
.
−
P
A_n=M(1+r)^n-P(1+r)^{n-1}-P(1+r)^{n-2}-.....-P
An=M(1+r)n−P(1+r)n−1−P(1+r)n−2−.....−P
This above expression is the one of geometric progression, the sum of a geometric progression can be detrived by applying the following equation:
So in this case we don’t have the first term, so set that to 1, but we have the increasing factor which is (1+r). Therefore, the sum of
A
n
A_n
An can be computed by
M ( 1 + r ) n − P ( ( 1 + r ) n − 1 ) ⋅ r M(1+r)^n -P((1+r)^n -1)\cdot r M(1+r)n−P((1+r)n−1)⋅r
However, all the steps we have gone through are for A n A_n An which how many a student left. We are expecting a formula for monthly payment.Therefore, we should rework the formula:
P = M ( 1 + r ) n ( 1 + r ) − 1 \frac{M(1+r)^n}{(1+r)-1} (1+r)−1M(1+r)n
3.Conclusions
- system analysis is vital. It is helpful for us to understand what the instances in the system and what relationships may have among them.
- When dealing with the time-based question, we tend to manually calculate the expressions for the first several points to see if we can find pattern that can express the entire system.
- Always simplify the expressions.