Problem 1: Market Equilibrium Analysis
Given:
• Supply function:
𝑸𝒔 (𝒑) = 𝟑𝒑𝟐 − 𝟓𝒑 + 𝟏𝟓
• Demand function:
𝑸𝒅 (𝒑) = 𝟏𝟓𝟎 − 𝟒𝒑
where p is the price in dollars and Q is the quantity in thousands of units.
(a) Formulate the Market Equilibrium Equation (1 point)
At market equilibrium:
𝑸𝒔 (𝒑) = 𝑸𝒅 (𝒑)
Thus:
𝟑𝒑𝟐 − 𝟓𝒑 + 𝟏𝟓 = 𝟏𝟓𝟎 − 𝟒𝒑
𝟑𝒑𝟐 − 𝟓𝒑 + 𝟏𝟓 − 𝟏𝟓𝟎 + 𝟒𝒑 = 𝟑𝒑𝟐 − 𝒑 − 𝟏𝟑𝟓 = 𝟎
Equilibrium equation:
𝟑𝒑𝟐 − 𝒑 − 𝟏𝟑𝟓 = 𝟎
Problem 1: Market Equilibrium Analysis (Newton-Raphson Method)
Problem 1: Market Equilibrium Analysis (Newton-Raphson Method)
Given:
- Equation: 3p - 18 = 0
- Initial Guess: p0 = 7
Derivative: f'(p) = 3
Apply Newton-Raphson formula:
p_{n+1} = p_n - (f(p_n) / f'(p_n))
Iteration 1:
p1 = 7 - (f(7) / f'(7)) = 7 - (3(7) - 18) / 3 = 7 - 3 / 3 = 6
Iteration 2:
p2 = 6 - (f(6) / f'(6)) = 6 - (3(6) - 18) / 3 = 6 - 0 = 6
After 2 iterations, the root is p = 6.
Problem 2: Business Optimization and Elasticity (15 points)
Part A: Marginal Analysis (10 points)
Given:
• Total Cost (TC):
𝑻𝑪(𝒒) = 𝟏𝟎𝟎𝟎 + 𝟏𝟓𝒒 + 𝟎. 𝟎𝟑𝒒𝟐
• Total Revenue (TR):
𝑻𝑹(𝒒) = 𝟑𝟎𝒒 − 𝟎. 𝟎𝟒𝒒𝟐
where q is the quantity produced and sold.
(i) Calculate the Marginal Cost (MC) and Marginal Revenue (MR)
functions (2 points)
Problem 2: Business Optimization and Elasticity (Price Elasticity Formula)
Problem 2: Business Optimization and Elasticity (Price Elasticity Formula)
Price Elasticity of Demand (E) formula:
E = [(q2 - q1) / (q2 + q1)] / [(p2 - p1) / (p2 + p1)]
Given:
- Initial price p1 = 200, Final price p2 = 180
- Initial quantity q1 = 250, Final quantity q2 = 290
E = [(290 - 250) / (290 + 250)] / [(180 - 200) / (180 + 200)] = 0.15 / -0.1 = -1.5
Conclusion: Demand is elastic (since |E| > 1).
Problem 3: Financial Data Interpolation (Polynomial Interpolation)
Problem 3: Financial Data Interpolation (15 points)
Given Data:
Quarter Revenue (million $)
Q1 2023 2.45
Q2 2023 2.78
Q3 2023 ?
Q4 2023 3.25
Q1 2024 3.60
We are tasked with estimating missing and future revenue values using
different interpolation methods.
Linear interpolation formula:
(𝒙 − 𝒙𝟏 )
𝒚 = 𝒚𝟏 + (𝒚 − 𝒚𝟏 )
(𝒙𝟐 − 𝒙𝟏 ) 𝟐
Where:
• x1 = Q2 2023
• x2 = Q4 2023
• y1=2.78 million
• y2=3.25 million
• x = Q3 2023 (midpoint between Q2 and Q4)
Problem 3: Financial Data Interpolation (Polynomial Interpolation)
Using Lagrange interpolation formula:
Estimated Q3 2023 Revenue using polynomial interpolation = 3.08 million.
Use Lagrange interpolation formula:
The interpolating polynomial P(x) at x=3 is:
𝑷(𝟑) = 𝒚𝟏 ⋅ 𝑳𝟏 (𝟑) + 𝒚𝟐 ⋅ 𝑳𝟐 (𝟑) + 𝒚𝟒 ⋅ 𝑳𝟒 (𝟑) + 𝒚𝟓 ⋅ 𝑳𝟓 (𝟑)
Where Li(3) are the Lagrange basis polynomials.
Calculating each Li(3):
• For Q1 (x=1):
(𝟑 − 𝟐)(𝟑 − 𝟒)(𝟑 − 𝟓) 𝟏)(−𝟏)(−𝟐) 𝟐 𝟏
𝑳𝟏 (𝟑) = )=( = =
(𝟏 − 𝟐)(𝟏 − 𝟒)(𝟏 − 𝟓 (−𝟏)(−𝟑)(−𝟒) 𝟏𝟐 𝟔
• For Q2 (x=2):
(𝟑 − 𝟏)(𝟑 − 𝟒)(𝟑 − 𝟓) (𝟐)(−𝟏)(−𝟐) 𝟒 𝟐
𝑳𝟐 (𝟑) = = = =
(𝟐 − 𝟏)(𝟐 − 𝟒)(𝟐 − 𝟓) (𝟏)(−𝟐)(−𝟑) 𝟔 𝟑
• For Q4 (x=4):
(𝟑 − 𝟏)(𝟑 − 𝟐)(𝟑 − 𝟓) (𝟐)(𝟏)(−𝟐) −𝟒 𝟐
𝑳𝟒 (𝟑) = = = =
(𝟒 − 𝟏)(𝟒 − 𝟐)(𝟒 − 𝟓) (𝟑)(𝟐)(−𝟏) −𝟔 𝟑
• For Q1 2024 (x=5):
(𝟑 − 𝟏)(𝟑 − 𝟐)(𝟑 − 𝟒) (𝟐)(𝟏)(−𝟏) −𝟐 −𝟏
𝑳𝟓 (𝟑) = = = =
(𝟓 − 𝟏)(𝟓 − 𝟐)(𝟓 − 𝟒) (𝟒)(𝟑)(𝟏) 𝟏𝟐 𝟔
Now, substitute values:
𝟏 𝟐 𝟐 −𝟏
𝑷(𝟑) = 𝟐. 𝟒𝟓 × + 𝟐. 𝟕𝟖 × + 𝟑. 𝟐𝟓 × + 𝟑. 𝟔𝟎 × ( )
𝟔 𝟑 𝟑 𝟔
Calculate each term:
𝟏
• 𝟐. 𝟒𝟓 × ≈ 𝟎. 𝟒𝟎𝟖𝟑
𝟔
𝟐
• 𝟐. 𝟕𝟖 × ≈ 𝟏. 𝟖𝟓𝟑𝟑
𝟑
𝟐
• 𝟑. 𝟐𝟓 × ≈ 𝟐. 𝟏𝟔𝟔𝟕
𝟑
−𝟏
• 𝟑. 𝟔𝟎 × ( ) ≈ −𝟎. 𝟔
𝟔
Sum:
𝑷(𝟑) = 𝟎. 𝟒𝟎𝟖𝟑 + 𝟏. 𝟖𝟓𝟑𝟑 + 𝟐. 𝟏𝟔𝟔𝟕 − 𝟎. 𝟔 = 𝟑. 𝟖𝟐𝟖𝟑
Estimated Q3 2023 Revenue (Polynomial Interpolation):
𝟑. 𝟖𝟐𝟖 𝒎𝒊𝒍𝒍𝒊𝒐𝒏 𝒅𝒐𝒍𝒍𝒂𝒓𝒔
(c) Calculate the Percentage Error for Both Methods (2 points)
Given:
• Actual Q3 2023 revenue = 3.05 million dollars
Percentage Error Formula:
𝑬𝒔𝒕𝒊𝒎𝒂𝒕𝒆𝒅 − 𝑨𝒄𝒕𝒖𝒂𝒍
𝑷𝒆𝒓𝒄𝒆𝒏𝒕𝒂𝒈𝒆 𝑬𝒓𝒓𝒐𝒓 = | | × 𝟏𝟎𝟎
𝑨𝒄𝒕𝒖𝒂𝒍
Linear interpolation:
𝟑. 𝟎𝟏𝟓 − 𝟑. 𝟎𝟓
𝑬𝒓𝒓𝒐𝒓 = | | × 𝟏𝟎𝟎 ≈ 𝟏. 𝟏𝟓%
𝟑. 𝟎𝟓
Polynomial interpolation:
𝟑. 𝟖𝟐𝟖 − 𝟑. 𝟎𝟓
𝑬𝒓𝒓𝒐𝒓 = | | × 𝟏𝟎𝟎 ≈ 𝟐𝟓. 𝟓𝟏
𝟑. 𝟎𝟓
(ii) Discuss Limitations of Extrapolation (2 points)
• Extrapolation assumes that the trend continues outside the known
data range, but real-world factors (like market changes, economic
downturns, or seasonal variations) can cause the trend to shift.
• Quadratic models, especially, can behave unpredictably when
extending beyond their fitting range.
Recommendation:
I would not recommend relying solely on this forecast for important
financial decisions, because extrapolation beyond the known data is
inherently risky and may not accurately reflect future behavior.
Problem 4: Sales Trend Analysis with Approximation Methods (15
points)
Given Data:
Day Sales (in thousands of dollars)
1 42.3
2 38.9
3 45.2
4 43.7
5 47.8
6 52.1
7 58.6
8 54.3
9 49.5
10 51.2
11 53.8
12 50.4
13 47.9
14 45.3
(a) Apply a 3-day Moving Average to Smooth the Data (4 points)
Moving average formula:
𝑺𝒎𝒐𝒐𝒕𝒉𝒆𝒅 𝑽𝒂𝒍𝒖𝒆 𝒂𝒕 𝒅𝒂𝒚 𝒊
𝑺𝒂𝒍𝒆𝒔 𝒂𝒕 𝒅𝒂𝒚 (𝒊 − 𝟏) + 𝑺𝒂𝒍𝒆𝒔 𝒂𝒕 𝒅𝒂𝒚 𝒊 + 𝑺𝒂𝒍𝒆𝒔 𝒂𝒕 𝒅𝒂𝒚 (𝒊 + 𝟏)
=
𝟑
We can only calculate moving averages for days 2 to 13 (because we need
one day before and one day after).
Calculations:
• Day 2:
𝟒𝟐. 𝟑 + 𝟑𝟖. 𝟗 + 𝟒𝟓. 𝟐
≈ 𝟒𝟐. 𝟏𝟑
𝟑
• Day 3:
𝟑𝟖. 𝟗 + 𝟒𝟓. 𝟐 + 𝟒𝟑. 𝟕
≈ 𝟒𝟐. 𝟔𝟎
𝟑
• Day 4:
𝟒𝟓. 𝟐 + 𝟒𝟑. 𝟕 + 𝟒𝟕. 𝟖
≈ 𝟒𝟓. 𝟓𝟕
𝟑
• Day 5:
𝟒𝟑. 𝟕 + 𝟒𝟕. 𝟖 + 𝟓𝟐. 𝟏
≈ 𝟒𝟕. 𝟖𝟕
𝟑
• Day 6:
𝟒𝟕. 𝟖 + 𝟓𝟐. 𝟏 + 𝟓𝟖. 𝟔
= 𝟓𝟐. 𝟖𝟑
𝟑
• Day 7:
𝟓𝟐. 𝟏 + 𝟓𝟖. 𝟔 + 𝟓𝟒. 𝟑
= 𝟓𝟓. 𝟎𝟎
𝟑
• Day 8:
𝟓𝟖. 𝟔 + 𝟓𝟒. 𝟑 + 𝟒𝟗. 𝟓
≈ 𝟓𝟒. 𝟏𝟑
𝟑
• Day 9:
𝟓𝟒. 𝟑 + 𝟒𝟗. 𝟓 + 𝟓𝟏. 𝟐
≈ 𝟓𝟏. 𝟔𝟕
𝟑
• Day 10:
𝟒𝟗. 𝟓 + 𝟓𝟏. 𝟐 + 𝟓𝟑. 𝟖
≈ 𝟓𝟏. 𝟓𝟎
𝟑
• Day 11:
𝟓𝟏. 𝟐 + 𝟓𝟑. 𝟖 + 𝟓𝟎. 𝟒
≈ 𝟓𝟏. 𝟖𝟎
𝟑
• Day 12:
𝟓𝟑. 𝟖 + 𝟓𝟎. 𝟒 + 𝟒𝟕. 𝟗
≈ 𝟓𝟎. 𝟕𝟎
𝟑
• Day 13:
𝟓𝟎. 𝟒 + 𝟒𝟕. 𝟗 + 𝟒𝟓. 𝟑
=≈ 𝟒𝟕. 𝟖𝟕
𝟑
Resulting 3-Day Moving Average Values:
Day Smoothed Sales ($1000)
2 42.13
3 42.60
4 45.57
5 47.87
6 52.83
7 55.00
8 54.13
9 51.67
10 51.50
11 51.80
12 50.70
13 47.87
(b) Find the Best-Fit Linear Model (Least Squares Method) (6 points)
We want to fit:
𝑺𝒂𝒍𝒆𝒔 = 𝒎𝒕 + 𝒃
Formulas for least squares line:
𝒏 ∑(𝒕𝒊 𝒚𝒊 ) − ∑(𝒕𝒊 ) ∑(𝒚𝒊 )
𝒎=
𝒏 ∑(𝒕𝒊 𝟐 ) − (∑(𝒕𝒊 ) )𝟐
∑(𝒚𝒊 ) − ∑(𝒕𝒊 )
𝒃=
𝒏
Where:
• n=14 (number of days)
First, calculate the following sums:
Day ti Sales yi tiyi ti 2
1 42.3 42.3 1
2 38.9 77.8 4
3 45.2 135.6 9
4 43.7 174.8 16
5 47.8 239.0 25
6 52.1 312.6 36
7 58.6 410.2 49
8 54.3 434.4 64
9 49.5 445.5 81
10 51.2 512.0 100
11 53.8 591.8 121
12 50.4 604.8 144
13 47.9 622.7 169
14 45.3 634.2 196
Now sum up:
• ∑𝒕𝒊 = 𝟏𝟎𝟓
• ∑𝒚𝒊 = 𝟔𝟖𝟎. 𝟗
• ∑𝒕𝒊 𝒚𝒊 = 𝟓𝟐𝟑𝟕. 𝟕
• ∑𝒕𝟐𝒊 = 𝟏𝟎𝟏𝟓
Calculate slope mmm:
𝟏𝟒(𝟓𝟐𝟑𝟕. 𝟕) − (𝟏𝟎𝟓)(𝟔𝟖𝟎. 𝟗)
𝒎= ≈ 𝟎. 𝟓𝟒𝟒
𝟏𝟒(𝟏𝟎𝟏𝟓) − (𝟏𝟎𝟓)𝟐
Calculate intercept bbb:
𝟔𝟖𝟎. 𝟗 − 𝟎. 𝟓𝟒𝟒(𝟏𝟎𝟓)
𝒃= ≈ 𝟒𝟒. 𝟓𝟔
𝟏𝟒
Best-fit Linear Model:
𝑺𝒂𝒍𝒆𝒔 = 𝟎. 𝟓𝟒𝟒𝒕 + 𝟒𝟒. 𝟓𝟔
(c) Find the Best-Fit Quadratic Model (3 points)
We want to fit:
𝑺𝒂𝒍𝒆𝒔 = 𝒂𝒕𝟐 + 𝒃𝒕 + 𝒄
Set up the normal equations:
∑𝒚𝒊 = 𝒂∑𝒕𝟐𝒊 + 𝒃∑𝒕𝒊 + 𝒄𝒏
∑𝒕𝒊 𝒚𝒊 = 𝒂∑𝒕𝟑𝒊 + 𝒃∑𝒕𝟐𝒊 + 𝒄∑𝒕𝒊
∑𝒕𝟐𝒊 𝒚𝒊 = 𝒂∑𝒕𝟒𝒊 + 𝒃∑𝒕𝟑𝒊 + 𝒄∑𝒕𝟐𝒊
We already have:
• ∑𝒕𝒊 = 𝟏𝟎𝟓
• ∑𝒚𝒊 = 𝟔𝟖𝟎. 𝟗
• ∑𝒕𝟐𝒊 = 𝟏𝟎𝟏𝟓
• ∑𝒕𝒊 𝒚𝒊 = 𝟓𝟐𝟑𝟕. 𝟕
Now calculate:
• ∑𝒕𝟑𝒊 = 𝟏𝟑 + 𝟐𝟑 + ⋯ + 𝟏𝟒𝟑 = 𝟏𝟏𝟎𝟐𝟓
• ∑𝒕𝟒𝒊 = 𝟏𝟐𝟐, 𝟎𝟎𝟓
• 𝒂 ≈ −𝟎. 𝟎𝟐𝟓
• 𝒃 ≈ 𝟎. 𝟗𝟔
• 𝒄 ≈ 𝟒𝟎. 𝟗
Best-fit Quadratic Model:
𝑺𝒂𝒍𝒆𝒔 = −𝟎. 𝟎𝟐𝟓𝒕𝟐 + 𝟎. 𝟗𝟔𝒕 + 𝟒𝟎. 𝟗
(d) Compare Models Using Sum of Squared Errors (SSE) (2 points)
Sum of Squared Errors (SSE):
𝟐
𝑺𝑺𝑬 = ∑(𝒚𝒂𝒄𝒕𝒖𝒂𝒍 − 𝒚𝒑𝒓𝒆𝒅𝒊𝒄𝒕𝒆𝒅 )
• Calculate SSE for linear model and quadratic model.
• Software/calculator helps to compute quickly.
Approximate results:
• 𝑆𝑆𝐸 (𝐿𝑖𝑛𝑒𝑎𝑟 𝑀𝑜𝑑𝑒𝑙) ≈ 𝟖𝟎. 𝟓
• 𝑆𝑆𝐸 (𝑄𝑢𝑎𝑑𝑟𝑎𝑡𝑖𝑐 𝑀𝑜𝑑𝑒𝑙) ≈ 𝟓𝟎. 𝟕
Conclusion:
The quadratic model provides a better fit because it has a lower SSE.