Вы находитесь на странице: 1из 5

Function

Binary
Variables
(with
solver)
Binding
and Nonbinding
variables
COUNTIF

Decision
Tree
Derivativ
e
Distributi
on Costs
Dummy
Variable
Game
Theory
Goal
Seek
HardCoding
Hypothe
sis
testing
IF Stmts

Linear
Regressi
on

MIN
FUNCTIO
N
MulitLinear
Regressi
on
NORM
DIST
NORM
DIST for
greatest
EV
NORMIN
V

Objectiv
e Coefficient
Objectiv
e
Function
One way
Table
OR
function
OR with
an IF
STMNT
P-VALUE

R
(correlati
on coefficient)

R^2
(coefficient
of
determin
ation)
RAND ()
+ IF
stateme
nt +
Oneway
Table
RAND()

RAND()
in an IF
STATEME
NT

Reduced
Cost
Shadow
Price
SOLVER

Standard
Error
Tables
with IF
STMNTS
T-Test

Two Way
Tables

Rules
Linear vs. Non Linear: linear must have only addition, subtraction or
multiplication. Where are non linear is max min, if statements, division, and cross
product.
A linear programming problem (LP) is an optimization problem for which we do
the following:
1) We attempt to maximize (or minimize) a linear function of the decision variables.
The function that is to be maximized or minimized is called the objective function.
2) The values of the decision variables must satisfy a set of constraints. Each
constraint must be a linear equation or linear inequality. Linear meaning mx+b, not
mxy/b, not m^b, not min(xy)+b, just + or x times a number.
One Way vs Drag Down: one way will do the same thing as drag down, but
instead of dragging five, you can put the one formula in the corner and itll plug the
rest. The catch, if you want to do solver you cannot do a one way.
Skewed Graphs: skewed to the right makes mean greater than the median, when
the mean is greater than the median, then we know the percentile of the mean is
higher than 50%(50th percentile being the median)
Skewed to the Left makes the mean smaller than the median, when the mean is
smaller than the median, then we know the percentile of the mean is lower than
50%
Normal Distribution, no skews, mean and median should be close to the same
Making a Conceptual Solver Model:
1) State all variables, x1=amount allocated to location 1 and plant
1....x12=amount allocated to location 6, plant 2 (make sure every variable is

covered, if your decision variable chart is 6*6, then you should have 36
variables)
2) What is your objective function? Max profit=rev cost
a. How do you get rev? (break down)
b. How do you get costs (a+b+c)
c. How do you get a, then b, then c, until there can be no questions on
how any part of the objective function came to be
3) Then list your constraints
a. Capacity
i. Is it per hour>/60, does not specify time, than assume no time
increment and don't make units into time units
b. Environmental
i. Is it WEIGHTED AVERAGE of two variables?
1. (#*(h/h+s))+(2#*(s/s+h))
c. Demand
i. Can you handle more demand then capacity? NO.
ii. Is it profitable to ship more than demand? Probably not
d. Non negativity?
i. What numbers cant be negative? Demand, price,
e. Binary
i. Which element is binary, why is it binary? Do you have a yes or
no decision?
Constraint Rules
Type
formula
Atleast x amount
Do not include A if B is
included.
A can only be included if
B is included
A and B must both be
included, or not at all
A and B and C cannot
occur at the same time
A must happen if B and C
occur
Non-Negativity
Binary Variables
There are two types of
inventory, one worth 5,
the other worth 8; you
need a minimum of 7 in
order to have a
satisfactory product.

Binary
Binary

A+B+C+D+E>=5
A<=B-1

Binary

A-B<=0

Binary

A-B=0

Binary

A+B+C<3

Binary

A+1>=B+C

Regular or binary
Binary
Regular, continuous

A,b,c,d,>=0
Must be a one or zero
Non linear:
(5a+8b)/a+b>=7
Linear: CANT USE
DIVISION,
CROSSPRODUCT, MIN,
MAX, OR IF STATEMENTS
IN CONSTRAINT
(5a+8b)-7*(a+b)>=0

Вам также может понравиться