OIS 3660 FINAL STUDY
If interarrival times are fixed (IE deterministic), then the coefficient of variation equals
0
How is the setting of overbooking flights a Newsvendor setting?
Because airlines weigh the costs of overbooking versus empty seats
What do you think is the best reason for the quality loss function to be quadratic in nature?
Because the cost of low quality should increase more than linearly the further away we get from the desired quality
If we want to achieve mass customization, then we should really try to:
Bring the fan-out point as close as possible to the push-pull point
Annual turnover rate of inventory
COGS / Average Inventory
Which of the these quality dimensions lends itself the easiest to statistical measurement & control?
Conformance
Kaizen
Continuous improvement that involves all participants. Japanese term for continuous improvement
What two types of costs does the Newsvendor model weigh in making a decision regarding how much to produce/purchase?
Cost of underage and cost of overage Newsvendor is all about making the trade-off between having too much (cost of overage) and not having enough (cost of underage). When these two are the same, then order/produce the expected demand. When the cost of overage is higher, then order less than the expected demand (better to not have too much!), and when the cost of underage is higher, then order more than the expected demand (not so bad if there are leftovers and better to not run out!).
What should be considered when a process is NOT in control?
Determine what the cause of not being in control was. Whether long-term or temporary, deviations from control can reduce your quality, and thereby revenues and profits. By carefully monitoring and manipulating your process to fit control limits and guidelines, you can save waste and increase throughput.
With CPK, is higher or lower better
Higher. A Cpk of less than 1.0 means that the process is not capable of meeting its requirements.
What does centering the mean do to capability index of a process?
Increases the Cpk index
Which corner in the OM Triangle should you get close to according to Lean Operations?
Information
A process has a setup time of 10 mins (every batch or lot of items incurs 1 setup). Which lot size will result in the lowest implied utilization?
Lot size of 10 (largest size) The resource must spend time doing a setup once for every lot. So the larger the lot size, the fewer setups the resource needs to perform. Not doing a setup "frees up" the resource, that is, reduces its implied utilization. Thus a larger lot size results in the lowest implied utilization.
Meal service to the economy class on an airplane should be located where in the OM Triangle?
Low Capacity, medium Inventory, and medium Information
Different product types require different:
Management tasks; competitive modes. See powerpoint for more info
MAD
Mean Absolute Deviation
MAPE
Mean Absolute Percentage Error
If the cost of overage is higher than the cost of underage, then you should...
Order fewer than the average expected demand
For which of the following settings/scenarios would the cost constant (i.e., for the quality loss function) be the HIGHEST?
Pacemaker chips. If something like bicycle tires or tennis balls were off specification people wouldn't even notice
What is the output of an airline?
Passengers at their destination
Poka-yoke
Safeguards built into a process to reduce the possibility of errors.
If forecasting errors are greater than 1, then
The MSE will be higher than the MAD. Since the MSE takes the average of the squared forecasting errors, it will be larger than the MAD which just takes the average of the forecasting errors, as long as the forecasting errors are larger than 1 (ie numbers smaller than 1 become even smaller when we square them, numbers larger than 1 become larger when we square them.) The answer most often given "The MSE will be more reliable has no meaning in this context, what does it mean for a measure of forecasting error to be more reliable?
When a time series is extremely stable, then which forecasting method is most appropriate?
The averaging method
If we have determined the Economic Order Quantity to be 1,342 units, but we decide to order 1,500 units every time, since this is an easier number to deal with, then which will be the result, as compared with the EOQ?
The holding costs will be higher
How to determine seasonal factor
The seasonal factor for a given season (or period or day) is determined by dividing that season's average by the overall average.
What does TIRP stand for?
Transparency, Incentive compatibility, Robustness, and Pinch-Point
What achieves reduction in the variability of interarrival times?
Turning away peak demand, scheduling appointments (to smooth out the demand), set a constant takt time (about the rate of output so that smooths rate of input). Using a push system wouldn't help as that doesn't influence the arrival rate in any way (just means as soon as the customer arrives we take it through all the steps of the process)
If the processing times in Littlefield are 0.1 day for each of the four process steps, and the specified lead time is 0.5 days, and we are dealing with quite some variability both in demand and in processing times, then:
We need to make sure we have enough capacity to ensure we can meet a 0.5 day throughout time (it is possible since the sum of the process times at the 4 process steps is 0.4 days).
Why does double marginalization lead to lower sales?
When all parties in the supply chain add a margin, the final price is higher, so fewer people are willing to buy the product/service. Effects are higher customer price, lower sales, lower supply chain profits
Interarrival time (a)
a = 1 / R If 10 jobs arrive per hour, then (on average) one job arrives every 6 min
A chef at Benihana can be best described as
an entertainer
Utilization higher than 100% is
implied utilization
When determining optimal price: when the demand curve is linear .....
it is optimal for a firm to price half-way between its cost and the y-axis intercept
Which requires the most skilled labor?
job shop
# of processors
m
Preponement
moving push-pull point upstream to fan-out point or holding unfinished state [before fan-out] of product until we receive order
job arrival rate
r
Bullwhip effect
the phenomenon in supply chains whereby ordering patterns experience increasing variance as you proceed upstream in the chain. Variability increases in the supply chain when moving from the customer "upstream"
MSE - Mean Squared Error
weights errors according to their squared values
Pure process time
the time to process one item on its own, not counting the setup time
After a careful analysis, managers at Haremon's have concluded that 20 customers per hour need to check out. If the sole cashier can check out 25 customers per hour, what is the cashier's (implied) utilization?
80% R = 20 customers/hr m = 1 c = 25/hour u = r / mc = 20 per hr / (1 * 25 per hr) = .80 or 80%
A process can be said to be capable when:
Cpk is greater than 1.5. All the data points fall well within the specification limits with a normal distribution. A process where almost all the measurements fall inside the specification limits is deemed a capable process.
MSE
Mean Squared Error
Why study ops?
It teaches you how to think logically, make money, and manage transformations from inputs to outputs.
What does "x-double-bar" mean?
The average of the averages
Job shop
A flexible process structure for products that require different inputs and have different flows through the process. Can make "whatever job you like", but it will cost you a lot.
Moving average forecasting method
A forecasting method that predicts that the next value will be the average of the last realized values.
If a process step has a LARGE setup time and a SMALL pure-processing time per unit, what batch size will result in the fastest throughput time for any ONE unit through this step? Assume there are multiple resources at this step, each with plenty of capacity.
A small batch. With multiple resources and plenty of capacity, the smaller batches can be processed simultaneously. (Again, for practical purposes, the emphasis on LARGE and SMALL is irrelevant as long as there is PLENTY capacity — the PLENTY is important to avoid queueing effects.)
The difference between 99.5% and 99.95% successful on a daily basis is...
HUGE! On a daily basis that is a small difference (0.45%), but on a yearly basis, this means: 99.95%^365 = 83.31% successful 99.5%^365 = 16.04% successful That is HUGE!
Linear regression equation
the equation for the line of best fit
What's the difference between a trend and seasonality?
A trend is a relatively steady increase (or decrease) over time, seasonality is a pattern that repeats (e.g., every year or even every week).
constant takt time
takt time refers to the amount of time a manufacturer has per unit to produce enough goods to fulfil customer demand. Frequently used within lean production lines, takt time is an essential tool in ensuring that goods flow through each build station in the most efficient manner.
What is the cause of Double Marginalization?
Myopic (nearsighted/lacking foresight or a realistic view) profit maximization
Let's assume that the x-bar chart and the R-chart show us that the process is in control. Does this mean that the process is also capable?
No: the means could all be very close together, but off target (e.g., all too large)
Which Lean Tool is most applicable if you have identified a process that should be improved and you would like to do so by standardizing the steps involved in the process?
SOP (Standard Operating Procedure) It is the process of documenting and standardizing processes for best practice and universal use
Product/process matrix
Categorizes processes into structures based on output volume and variety. Job shop is the most flexible, then batch flow, line flow, and flow shop.Example: french fries.If you cut them from potatoes and fry yourself, you are a job shop. If you go to a fancy dinner, they likely make them in batches. At McDonald's, they may make them in assembly line flow (frying in small batches). At the plant making the McCain frozen fries, they cut and package via continuous flow.
When a process has a capability index of 1, it is said to be able to meet the specification limits in 99.74% of the time. If a process is currently meeting the specification limits in 99.95% of the time, then the capability index must be:
Larger than 1
If we are dealing with a very unstable dataset, then which forecasting method will most likely result in the best (i.e., lowest) MAD, MSE, MAPE?
Last Value
PH produces many electronics products for the USA, Europe, and the UK. The plugs required for each of these geographical areas are different, so originally PH had produced three different versions of each product. Recently, they decided to create one version, and then ship these with three different plugs (i.e., of which two are redundant for the customer). What is the most likely result of this decision?
Lower safety stock costs, higher production costs
Providing idiosyncratic discounts to your customers can lead to a phenomenon called "forward buying". In the case of The Office, the customer Blue Cross will probably purchase a lot more paper next time they order, since they got such a great (and unexpected) discount. This would then force Dunder Mifflin to order more from their supplier etc., potentially even causing the Bullwhip effect... What is the best strategy to avoid forward buying?
Offer every day low pricing
What is the main input for an airline?
Passengers. Passengers are the FLOW UNITS for the airline process. The Input or Flow Unit is the thing going into the process that needs to be processed (changed, or worked on). For an airline, passengers are the things that the airline changes (move to another location.)
What is not a resource for an airline? Passengers, crew, airplanes, customs agents
Passengers. Passengers are the flow units. A resource is a part of the process that helps process the jobs, i.e. a machine or in this case airplanes, crew, and customs agents.
A linear regression was used to determine the linear relationship between the sales of Gulp Milk and the average price for milk across its competitors. If the average competitor price rises or falls, but all other factors remain equal, including the price of Gulp Milk itself, what kind of slope coefficient do we expect to find for this relationship?
Positive - when the average competitor price increases, Gulp Milk's price is relatively lower. Related to linear regression. If the average industry price (not including Gulps price) increases, then all other things equal, the sales for Gulp will increase (ie customers like a lower price, so if your competitors raises their prices then customers will switch to buying your product instead of the competitors products). Therefore, the coefficient would be positive. Conversely, if instead the question had asked about the coefficient related to Gulps own price (Ie if Gulps price was the independent variable) then the slope coefficient would be negative (ie if Gulp raises its price this will lead to lesser sales of Gulp milk.)
CPK
Process capability index that measures the potential for a process to generate defective outputs relative to either upper or lower specifications In simple words, it measures producer's capability to produce a product within customer's tolerance range.
A process can be said to be in control when:
The averages of the samples are all within the Lower and the Upper Control Limit. If you are within your upper and lower CONTROL limits you should be IN CONTROL
Mass customization is a bit of an oxymoron, but with the application of new technologies, it seems like we are getting closer to achieving it! What two things does mass customization try to achieve at the same time?
The benefits of mass production (low cost) and customization (high price)
To determine the average quality loss, we DO NOT need which of the following?
The characteristic of a certain output, we need the average output instead
Which is harder to determine, the cost of overage or the cost of underage?
The cost of underage
Most often, the MAD, MSE, and MAPE all point to the same forecasting method as the best one. Of they would all point to a different "best" forecasting method, then which forecasting method should we choose?
The one with the lowest MSE
If we increased our order quantities such that we order less frequently, then which of the following would NOT occur?
The ordering costs increase
A process capability index (Cpk) of 1 indicates that
The process is capable at the 3 sigma level. The sigma level can vary but 3 sigma is fairly standard and the mode for this class.
X-bar chart
A control chart used to monitor changes in the mean value of a process. The X-bar chart shows how the mean or average changes over time.
A production process stage takes a generic product and differentiates this into several different versions. When this happens, this is called:
A fan-out point. A fan-out point is where a process separates (or "fans out") into multiple product types or customizations.
Through the application of control charts, we are trying to detect which of the following?
Assignable cause variation
Where does Mass Customization fit in the Product/Process Matrix from Module 1?
At high volume and high variety (i.e., blending the benefits of the flow shop with those of the job shop)
How many different types of machines are there in the Littlefield Simulation? (How many stations are there in the game?)
3
Postponement
the strategy of the delay of any modifications or customization (final activities) to a product as long as possible in the production process, likely until all the orders are received.
Which is harder to achieve, a process capability index of 1 or 2?
2
The differences between a job shop and a flow shop affect EVERYTHING in the organization, for example:
- Material flow (jumbled vs. linear);- Equipment (general vs. specialized);- Capacity utilization (fluctuating vs. consistently high), etc.
Flow Shop
A flow shop can make "only one thing" in volume (it is productive) but it won't cost you much because they make so much of it in bulk.
Basic type of inventory:
Raw materials, finished goods, Work-in-process
A party of 5 goes to the Green Iguana. The waiter takes 10 seconds to get from the kitchen to the table (or vice versa), and takes 30 seconds to manually take each person's order at the table. The waiter always starts in the kitchen, goes out to a table, takes the order, and then delivers the order to kitchen. If the waiter takes the table's orders in a batch of 5, what is the "pure process time"?
30 seconds.Each person requires 30 seconds of ordering time, regardless of setup time.
If a document requires several different signatures, and the signees are spread out all over the office building, this represents which of the following types of waste?
Transport
If a process step has a LARGE setup time and a SMALL pure-processing time per unit, what batch size will result in the fastest completion of ALL of the units at this step? Assume there is only ONE resource at this step.
A large batch. To complete ALL the units you need to complete the "last" unit. The emphasis on LARGE and SMALL in the question is irrelevant: if setup time > 0 this ALWAYS holds.
Eight Dimensions of Quality
1. Performance 2. Features 3. Reliability 4. Conformance 5. Durability 6. Serviceability 7. Aesthetics 8. Perceived Quality
Which of the following products is most likely produced in a flow shop? Nascar race cars, custom made bicycles, airplanes, crude oil
Crude oil. It is processed in thousands if not millions of gallons and all conforms to the same specifications.
Further commentary on takt time
- Lean ops lingo has influence of foreign words bc of the influence of Toyota and others (taktzeit is German) - variability is bad (in a production setting) so one concept of lean ops is to take variability out of the production stream. We want steady constant times between start of 1 job and start of the next job. Even if actual customer requests (ie arrivals) are random, we may produce in a constant steady fashion rather than produce in the random fashion that would track the actual customer arrivals. (meaning we will hold a bit of finished goods inventory or else make some customers wait a bit) - We still want to produce at a rate (R) equal to the demand rate, no more and no less than we can sell on average. So we want to set a constant takt time - In the ice cream scooping problem, demand rate is 100 custs per hour, so in our nomenclature R = 100 / hr. Said another way, we want to scoop out a serving every 36 seconds. We say the takt time is 36 seconds. - So if we scoop out a serving of ice cream every 36 seconds, we will be producing to the demand rate of 100/hr. Luckliy the customers are arriving uniformly, meaning they are also arriving every 36 seconds rather than arriving randomly (aka no melted ice cream.) if arriving randomly, we might not want to set a constant takt time!
Lean management includes:
- Reducing variability in arrivals by setting a constant takt time - Reducing process time by reducing transportation, motion, and overprocessing - Reducing variability in process times by following standard operating procedures - Reducing defects - Reducing inventory by reducing wait time and by reducing overproduction
If there is no lead time (i.e., an order is fulfilled instantaneously), then what would be the optimal reorder point?
0. ROP = safety stock + expected lead time demand, but both terms would be zero if the lead time is 0!
If a process step has a LARGE setup time and a SMALL pure-processing time per unit, what batch size will result in the highest utilization?
A small batch. Smaller batches means more setups, so this increases the utilization. The emphasis on LARGE and SMALL in the question is irrelevant. If setup time > 0 then this ALWAYS holds true.
If a process step has a SMALL setup time and a LARGE pure-processing time per unit, what batch size will result in the fastest throughput time for any ONE unit through this step? Assume there are multiple resources at this step, each with plenty of capacity.
A small batch. With multiple resources and plenty of capacity, the smaller batches can be processed simultaneously.
LSL, USL, CPK
LSL stands for Lower Specification Limit and USL stands for Upper Specification Limit. Often we describe CPK (process capability index) as the capability the process is achieving whether or not the mean is centered between the specification limits.
If you want to produce a standard product in extremely large quantities, then you should work according to a
Flow Shop. Flow shops specialize in mass-producing standard products at low cost. (Line flow is an assembly line and is used to produce in relatively larger quantities but is less automated than a flow shop.)
Utilization (u)
Fraction of time a resource is busy. Utilization cannot exceed 1 (i.e. 100%), if u > 100%, we call it implied utilization.If there is variability then u may need to be kept substantially below 100%. Utilization = u = R p / m
Product life cycle
In marketing it is growth, maturity, decline. In the growth phase, the product is made in a job shop, with growth, the product moves down the spectrum to flow shop. Tensions arise between flexibility and productivity. There are tradeoffs and we need to make choices.
Which of the four potential problems described as TIRP is the main cause for the difference in cost of providing healthcare between Al Paso and McAllen's?
Incentive (in)compatibility
Where does the firm get this product?
It takes some kind of input and transforms it into an output - the product.EX: Hospital: transforms sick patients into healthy patients. Restaurant: transforms hungry customers into satisfied customers.OM teaches you how to manage this transformation, think logically, and analyze strengths and weaknesses in a firm's operations in order to improve processes.
Imagine a simple process in which 5 ticket booths are selling tickets for an opera. If customers are coming at a rate of 15 / hour, and the utilization of the 5 booths is 50%, then how long does a customer spend at a booth (each customer goes to only one booth)?
Solve for p: u = Rp / m u = .50 = rp / m R = 15 per hour m = 5 Solving for p, p = 10 minutes
How to determine forecast based on moving average (without seasonality)
Take average of the last few quarters, amount is dependent on n. If n = 2, take average of last two quarters
If a process is best situated close to the information corner of the OM triangle, then this must mean:
There is little waiting time for this process, capacity is expensive so we limit how much we buy leading to high u, there is low variability in arrivals and processing times
Hammon's is contemplating opening a little stand-alone satellite store. After a careful analysis, managers have concluded that 25 customers per hour will need to check out at the satellite store (this will be the long-run average demand at the store). They are considering hiring two cashiers, each of whom can check out a customer in 5 minutes on average. What can you say about this system?
This system will be unworkable since the implied utilization is greater than 100%. The managers at Hammon's need to do something.
What are we trying to detect through statistical process control (SPC)?
Variation that we can assign a cause to and thus address
If a process is best situated close to the inventory corner of the OM triangle, then this must mean:
Wait or throughput times are long, capacity is very expensive so we limit how much we buy leading to high u, there is a lot of variability in arrivals and processing times.
To determine WIP Limit (Work in progress limit)
We need to apply Little's Law: I = RT, where R is the arrival/departure rate (12 jobs per day), and T is the (intended) throughput time (0.25 days), so I = 12 jobs per day * 0.25 days = 3 jobs
Double marginalization
When each firm in a vertical chain marks up its price above its marginal cost, thereby increasing the price of the final product. Myopic means "short-sighted" and in the supply chain setting it suggests each firm maximizes its own profit without considering the other firm(s) in the supply chain. As such, both (or more) firms add a margin to their costs (too large a margin, as it turns out), yielding two (or more) large margins, and hence the term double marginalization.
Can a process be in control but not capable?
Yes - for example when the averages of the samples are all very close together, but all outside the specification limits. Key words here are "outside specification limits." Cpk, or capability, is a measurement of your ability to meet customer specifications. As control and specification are two separate measurements, your data could conform to your own set of control data but not your customer's specifications, and therefore be in control but not capable.
Control R-chart
a type of control chart used to monitor the process variability (as the range) when measuring small subgroups (n ≤ 10) at regular intervals from a process. Each point on the chart represents the value of a subgroup range. The center line for each subgroup is the expected value of the range statistic.
High quality is important in:
all industries. Manufacturing, services, healthcare
If there are different resources, the system capacity is equal to the capacity of the slowest resource
also called the bottleneck
Capacity of a resource (c)
c = 1 / p If a resource has a processing time (p) of 10 min, the capacity (c) is 6/hour
Process time
p