Title: Review Simulation, Inventory Models, Forecasting
1ReviewSimulation, Inventory Models, Forecasting
- Dr. A. Kleinstein, May 2005
2Simulation
- Simulation Using computer to replicate the
characteristics of a real system. - When the probability distributions of the input
variable are known the computer can generate
values that have those probability distributions.
- Discrete Distribution In Excel use rand() and
the vlookup table. By hand, use a table of
random numbers. - The model may be expressed through a flow chart
that shows what is done with the input variables. - Program the model into the computer, or follow
the model with hand calculations. - Compute the relevant statistics.
- Examples Harrys Tire, Simpkins Hardware,
Three Hill Power Company, Port of New Orleans,
vlookup example, day 26.
3Linear Programming
- Linear Programming technique to maximize or
minimize a linear function where the variables
are subject to linear constraints. This means
that the functions that represent the use of
resources are linear functions, and that they are
, or to constants. (The constants represent
the total amount of resources available.) - In applying this to real world problems the
variables are the amount to produce, and we make
the assumptions of - certainty the coefficients of the objective and
constraint functions are known and do not change - proportionality the constraint equations
accurately reflect the use of resources. Thus
producing twice as much of a product uses twice
the resources, etc.
4Linear Programming, cont
- additivity the objective function accurately
reflects the objective. Thus, each product
contributes additively to the objective function. - divisibility the solution values that represent
the amount of product to produce make sense even
if they are fractional values. - non-negativity all variables must be positive.
- When formulating a linear programming problem, be
sure that the constraint equations are to the
left of the equality of inequality signs, and the
constants are to the right.
5Linear Programming, cont.
- Linear programming problems can be solved with
Excel using the solver. - When using the solver, put the values of the
coefficients and the objective function in rows
of cells, and starting values for the decision
variables in a row of cells. Express the linear
function using the Excel command sumproduct.. - The sensitivity report, or rerunning the solver
with new values for the resources available, can
tell how much the objective function will change
for a unit change in resources available. - Examples
- Chapter 7 Flair Furniture Company, Holiday Meal
Turkey Ranch. - Chapter 8 Media Selection, Marketing Research,
Portfolio Selection, Greenberg Motors. - Day 23 Portfolio Selection
6Inventory Models
- Inventory Models PowerPoint presentation, day
16. - Economic Order Quantity Model Used since 1915.
Limited in application, but this model is useful
as a conceptual tool, and starting off point. - Assumption
- demand is known and constant
- lead time is known and constant
- receipt of inventory is instantaneous
- quantity discounts are not possible
- only variable costs are the cost of placing an
order, the ordering cost, and the cost of holding
or storing inventory over time, the holding or
carrying cost. - Orders are placed so that stockouts and shortages
are completely avoided.
7Inventory Model, cont
- The minimum cost occurs when the carrying cost
equals the holding cost. This gives the
equations on page 197. EOQ is the amount to
order each time, so as to minimize holding plus
carrying cost. - If we allow quantity discounts, then we must
include the cost of purchasing the items. - total cost material cost ordering cost
carrying cost. - In the case when quantity discounts are
available, we often express the carrying cost as
a percentage (I) of the unit cost (C). - When quantity discounts are available, compute
the EOQ amount for each available price. See
quantity discount model steps on page 207. - Example
- Chapter 6 Sumco Pump Company. Brass Department
Store - Day 18 Appack example and Excel solution
8Forecasting
- Forecasting PowerPoint presentation, day 9.
- Forecasting method
- Moving average
- weighted moving average
- linear regression trend projections and casual
forecasting - Measures of accuracy
- Graphs visual measure
- MAD
- MSE
- SE standard error of the estimate. When using
linear regression for prediction, 95 of the time
the actual value is expected to be within 2
standard errors(2SE) of the predicted value.
9Forecasting, cont.
- Correlation coefficient when above 0.6 we
generally accept that the regression line should
be used for prediction. - When using Excel, the slope is computed with the
slope function, the y-intercept with the
intercept function, and the correlation
coefficient with the correl function. By hand,
see formulas on page 156, 157, and 169 - Examples
- Chapter 5 Wallace Garden Supply, Midwestern
Manufacturing Company, Triple A Construction
Company - Day 12, solved example.
10Review Models, Probability, Decision Theory
- Qant 305 Spring 2005
- Dr. A. Kleinstein
11Quantitative Analysis
- Mathematical tools have been used for thousands
of years - QA can be applied to a wide variety of problems
- One must understand the specific applicability
of the technique, its limitations and its
assumptions
12 Input/Process/Output
- Scientific Approach to Managerial Decision Making
- Consider both Quantitative and Qualitative Factors
Meaningful Information
Quantitative Analysis
Raw Data
13The QA Approach
Define the Problem
Develop a Model
Acquire Input Data
Develop a Solution
Test the Solution
Analyze the Results
Implement the Results
14Modeling in the Real World
- Models are complex
- Models can be expensive
- Models can be difficult to sell
- Models are used in the real world by real
organizations to solve real problems
15A Model Can be Mathematical Equations
Profit Model
Revenue (Price per Unit) ? (Quantity Sold)
Expenses Fixed Cost - (Variable Cost/Unit)
? (Quantity Sold)
16Breakeven Quantity Model
P price Q quantity sold F fixed cost V
variable cost/unit
Profit PQ-F-VQ
Solve for Q, the breakeven quantity by setting
profit 0 F PQ VQ, thus Q
F/(P V)
Breakeven Quantity F/(P-V)
17Mathematical Models Characterized by Risk
- Deterministic models - we know all values used in
the model with certainty - Probabilistic models - we know the probability
that parameters in the model will take on a
specific value
18Probability
- Life is uncertain!
- We must deal with risk!
- A probability is a numerical statement about the
likelihood that an event will occur - 0 ? P(event) ? 1
19Example Rolling a 6-sided Die
- Outcome
- of Roll
- 1
- 2
- 3
- 4
- 5
- 6
- Probability
- 1/6
- 1/6
- 1/6
- 1/6
- 1/6
- 1/6
- Total 1
Rolling a die has six possible outcomes
20Random Variables
- Discrete random variable - can assume only a
finite or limited set of numeric values- i.e.,
the number of automobiles sold in a year. - Continuous random variable - can assume any one
of an infinite set of numeric values - i.e.,
temperature, product lifetime.
21Discrete Random Variable
A random variable includes the possible outcomes
and the probability of getting those outcomes.
Probability Distribution
22Expected Value of a Discrete Random Variable
23Variance of a Discrete Random Variable
24Continuous Random Variable
25Area Under the Curve gives Probability
P(X
26Areas Under the Normal Curve
One, two, and three standard deviations from the
mean.
27Excel Functions for aNormal Random Variable
- Normdist(X,µ,s,true) P(x
- Norminv(prob, µ, s) X, where P(x
28Area Under the Normal Curve using Excel, Example 1
P(X ,true)
29Area Example 2
P(X rue)
30Area Example 3
P(110 Normdist(125,100,20,true) Normdist(110,100,20
,true)
31Decision Making Under Risk
- Risk means that you know the probability of
occurrence for each state of nature.
32Decision-Making Under Risk
- Clearly define the problem at hand
- List the possible alternatives
- Identify the possible outcomes (states of nature)
and probabilities of getting those outcomes. - List the payoff or profit of each combination of
alternatives and states of nature. - Compute the Expected Monetary Value (EMV), and
choose the alternative with the highest EMV.
33Decision-Making Under Risk
Choose the alternative with the highest Expected
Monetary Value (EMV)
34Example Three alternatives and Two States of
Nature
35Choose Alternative with the Highest EMV
Choose this alternative. Highest EMV.
36Perfect Information
- Perfect information means that you know the state
of nature that will occur. - You might know this because of market research.
- Knowing the state of nature does NOT alter the
probability that the state will occur. It only
means that you have determined which of the
states will occur before you choose the
alternative.
37Expected Value With Perfect Information (EV PI)
Example EV PI 200,000.50 0.50 100,000
38Expected Value of Perfect Information (EVPI)
- EVPI places an upper bound on what one would pay
for additional information - EVPI is the expected value with perfect
information minus the maximum EMV
39Calculating EVPI
- EVPI EV PI - max(EMV)
- Example
- EVPI 100,000 - 40,000 60,000
40Expected Opportunity Loss
- EOL is the cost of not picking the best solution
- EOL Expected Regret
- EOL best payoff in a column payoff in column
cell.
41Computing EOL - The Opportunity Loss Table
42Making a Decision with EOL Table Under Risk
- Calculate the EMV for each alternative in the EOL
table. - Choose the alternative with the minimum EMV.
- The value of the minimum EMV is also the EVPI.