Professional Documents
Culture Documents
SC0x M1Unit2 AnalyticsBasics ANNOTATED
SC0x M1Unit2 AnalyticsBasics ANNOTATED
ctl.mit.edu
Decision making is at the
core of supply chain management.
Increasing degree of
abstraction and speed
Increasing degree of
realism and cost
5
Source: Ragsdale, 2004
Roadmap for the Course
• Deterministic – Prescriptive Modeling
Basic functions & algebra
Classical optimization (calculus)
Math programming (LPs, IPs, MILPs, & Non-Linear)
• Stochastic/Uncertainty – Predictive & Descriptive
Basic probability and distributions
Statistical analysis (hypothesis testing)
Econometric modeling (regression)
Simulation
6
SCx Approach to Modeling
- Educating Drivers not Mechanics!
All images are CC0 Public Domain from https://pixabay.com and https://openclipart.org 7
Mathematical Functions
8
Mathematical Functions
“ . . . a relation between a set of inputs
and a set of permissible outputs with the
property that each input is related to
exactly one output.”
source: Wikipedia
we say:
y = f (x)
“f of x” or that “y is a function of x”
y = ax + b constant,
fixed value, or
independent y-intercept
slope or
variable
variable value
1
a
b
x
10
Examples: Linear Functions
• Truckload Transportation Costs:
cost = f(distance) = $200 + 1.35 $/km * (distance)
• Warehousing Costs
cost = f(# cases) = €2,500 + 2.5 €/case * (# cases)
• Profit Equation
profit = f(volume) = (r-c) * v + d
where:
r = revenue per item(¥/item)
c= cost per item (¥/item)
v = volume sold (items)
d = fixed cost (¥)
11
Quadratic Functions
12
a>0
Quadratic Functions y
This is the x
dependent
variable function, f(x).
y = ax + bx + c
a<0
2 y
x
independent constants
variable
- b ± b - 4ac
2
r1 ,r2 =
2a
14
Example: Finding Roots
(i) y = 2x2 so that a=2, b=c=0
- b ± b 2 - 4ac
0 0 2 4(2)(0) 0 r1 ,r2 =
r1 , r2 0 2a
2 ( 2) 4
(ii) y = 2x2 – 6x + 4 so that a=2, b=-6, c=4
16
Quadratic Functions in Practice
• Example: Manufacturing iWidgets – what price to set?
Cost of producing iWidgets is a linear y
function of the number produced, x: a 1
Maximum profit
Price at which we
maximize profit
Roots of quadratic
18
Quadratic Functions in Practice
• Example: Parcel Trucking – impact of weight
Parcel carriers combine many orders into a single shipment.
The cost of an individual order is a function of its weight, w.
However, it is not linear – it is tapering.
cost =f(weight) = 33 + 0.067w – 0.00005w2
€ 55
Cost Per Order (€) as a Function of Weight (kg)
€ 50
Cost Per Order (€)
€ 45
€ 40
€ 35
€ 30
0 50 100 150 200 250 300
Weight of Order (kgs)
19
Other Common Functional Forms
20
Power Function y=f(x) = axb
The shape of the curve is dictated by the value of b
3.00
if b<0, then function decreases
asymptotically to f(x)=0 and if b>1, then function
2.50
note that f(0) is undefined, increases exponentially
e.g., y=ax-1=a/x
2.00
1.50
0.50
if b=1, then the
function is linear
0.00
- 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00
21
Exponential Functions y=abx
80.00
0.00
- 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00
22
Logarithms
23
Other Inverse Relationships
Logarithms y=ax
y=a+x
x=y÷a
x=y-a
y=bx logb(y)=x
y is the value of b raised x is the power that I need to
to the xth power. raise the base, b, to equal y.
ln(2) = n ln(1.07)
Rearranging gives us: n = ln(2) / ln(1.07) = 0.693 / 0.182 = 10.24 = T
We could also use log10 where T = log(2)/log(1.07) = 10.24
The investment will double in value in 10.24 years.
70
26
Multivariate Functions
INPUT (x1, x2, . . . xn)
These are just functions with more than one
independent variable.
28
Properties of Functions
29
Properties of a Function: Convexity
A function is convex if it “holds water”
y y y
x x x
30
Properties of a Function: Continuity
A function is continuous if you can
draw it without lifting pen from paper!
y y y
x x
x
31
Key Points from Lesson
32
Key Points from Lesson (1/2)
• Different models used for different purposes
Descriptive – what has happened?
Predictive – what could happen?
Prescriptive – what should we do?
• Functions y=f(x)
Linear functions where y= ax + b
Quadratic functions where y= ax2 + bx + c
Power functions where y= axb
Exponential functions where y= abx
33
Key Points from Lesson (2/2)
• Logarithms
y=bx is equivalent to logb(y) = x
Natural log ln(y) = loge(y)
• Multivariate functions y=f(x1, x2, . . . xn)
Multiple inputs still lead to single output value
• Properties of functions
Convexity – does the function “hold water”?
Continuity – can I draw the function without
lifting my pencil
34
Questions, Comments, Suggestions?
Use the Discussion Forum!