COMM 204

2017W2
Final Review Package

By Jisoo Park
 Table of Contents

I. Process Analysis & Little’s Law

II. Multiple Types & Product Process Matrix
III. Variability in Processes: OM Triangle
IV. Queuing Analysis & Theory
V. Project Management
VI. Inventory Management
VII. Supply Chain Management
VIII. Quality Management
 I. Process Analysis & Little’s Law
Process Analysis
• Process flow diagram: Good tool for understanding & visualization of process
• Process mapping: Constructing the process flow diagram.
> Flow units: items flow through the process > Activities: transformation steps
> Resources & Time: perform activities > Buffers: storage flow unit

Process Analysis: Determine purpose Process mapping Capacity Analysis

i,e) Swim-Lane Flow Chart, Gantt Chart

• Flow time: length of time that a unit spends from the beginning to the end for specific “activity”
• Bottleneck: resource w/ lowest capacity rate in a process Determines the process capacity rate
• Capacity rate: maximum possible output rate
• Throughput rate: actual output rate (flow rate) Depends on the Capacity rate & Input rate
• Unit load: total amount of time the resource works to process
• Cycle time: average time between completion of units

Q. How can we increase the capacity rate of bottleneck?

1. Increase the size of resource pool 2. Decrease the unit load (t)

Little’s Law
• Little’s Law: relationship between average I, R, T

• A manager can influence any one of them by controlling other two

once two variables are chosen, third is determined
• Reduce flow time (T) by increasing throughput rate (R) and decreasing Inventory (I)
Utilization
• Utilization: gives us information about “excess capacity”

• Operational Challenge matching supply & demand

• Implied Utilization: captures excess demand in the short run allows us to capture the idea of OT

Inventory
• Inventory Build up

• Inventory Turnover
 II. Multiple Types & Product Process Matrix
Product Mix

• When multiple flow units go through a process, Product Mix needs to be considered.
Bottleneck depends on the Product mix.

• Flow Shop Vs. Job Shop

Small unit load for High Unit load for
each resource each resource
Labour Skills Low High
Equipment High Low
Specialization
Process Type Flow shop Job Shop

Product Process Matrix

• Processes take many different forms
• Evolution happens through the industry life cycle
 III. Variability in Processes: OM Triangle
Variability in Processes
• Variability comes from input and capacity rates
• Predictable variability & Unpredictable Variability

• Effect of Variability on Processes? AVERAGE

• Effect of Input variability with no buffer

Input variability can reduce the throughput rate

• Effect of Input variability with buffer

Build on Inventory Buffer, variability leads to increase in average Inventory in the process and
eventually increases the average flow time

The OM Triangle
➢ If a firm is striving to meet the Random D, then we can use
CIC as substitutions
➢ You cannot have low inventory, capacity, Information
acquisition at the same time
➢ “TRADE OFF”
 IV. Queuing Analysis & Theory
Queuing Analysis
• Pollaczek – Khinchin (P-K) Formula

• Impact of Utilization

Queuing Theories
• PK Formula Assumptions
- Single Server / queue / no limit on queuing length - Arrived people stays until get served. FIFO

1. G / G / 1 Queue 2. M / M / 1 Queue

3. M / D / 1 Queue 4. G / G / C Queue
- Multi-Server Queuing Model

- Risk Pooling: decreases queue length

dramatically. Demand Aggregation.

5. M / M / C Queue

Dealing with Variability

• Reduce it by improving information about input & Process (D & S)
• Manage it by choosing an appropriate buffer
• Adopt risk pooling
Queuing! XD !!
 V. Project Management
Project Management
• Project: set of related tasks & activities, directed towards some major output and requiring a
significant period of time to perform
- Often thought off as “one time “ occurrences, but often repeated with slight modifications
ex) construction, new product development, implementing new system, unique service activities…

• Product Process Matrix

• 1. Definition ( w/ project proposal) 2. Planning ( Gantt chart & Critical Path Diagram)
3. Execution ( Final product! ) 4. Retrospective ( Review & Feedbacks)
• Gantt Chart ?
• Critical Path Diagram

- Project duration: Use enumeration! Identify all paths between start and end node then find the
longest path
- To shorten the project duration, we need to shorten the longest path. In the longest path,
choose the cheapest task to shorten. If there are multiple longest paths, you need to find a set of
activities that will shorten all longest paths
• Why Critical Path? Comparing Critical Path and Bottleneck
- Determines project duration. All activities on the critical path are “Critical activities”
Any delays in critical activities will delay the whole project duration.

CRITICAL PATH BOTTLENECK

Determines project duration Determines process capacity
To shorten the duration To increase the capacity
May change when adding resources to critical May shift when adding resources
activities

Cost Information

• Crashing: an activity refers to reducing the time it takes to complete the activity
• Crash Time: minimum possible time to complete an activity ( NT - CT = Crushable Time)
• Crash Cost: Costs associated with crash time allow crashing with proportional cost
 V. Inventory Management
Forecasting
• Forecasting is vital: provide the basis for budgeting & cost control, relies on sales forecasting to
plan new products & promotions, to improve/ make decisions involving capacity planning, process
selection & inventory control. Basis for corporate long- Run planning
• Qualitative methods: executive judgement, market research, panel consensus, and Delphi method
• Quantitative methods: Time-Series Analysis, (weighted) moving averages, and exponential
smoothing
• Good Forecast?: smaller the errors, the better the forecast. Compute the MAD!
Forecast error = Forecast value – Actual value

• Forecasting Tips:
- Forecasts are always wrong.
- Long term forecasts are less accurate than short term forecasts.
- Aggregate forecasts are more accurate than disaggregate forecasts.

Inventory
• Inventory: the stock of any item/resource used in an organization
- Why should you hold inventory?:
Predictable & unpredictable variability, ES, transportation/flow times
- Why should you NOT?: Inventory increases certain costs, the sea of inventory hides problems

• Inventory Costs
- Ordering / Setup Costs (Fixed Costs): Fixed transportation Cost, Order processing Cost
- Holding Cost (Carrying Cost): Cost for storage, holding, insurance, working capital, etc
- Shortage Cost (Opportunity Cost): Lost sales, etc

• How much Inventory should you hold?

- Tradeoff #1: Inventory ordering cost vs. Holding cost EOQ Model
- Tradeoff #2: Cost of running out vs. Having excess inventory Newsvendor Model
Economic Order Quantity (EOQ)
• Assumptions: Known annual demand, constant demand rate, no uncertainty
• Notation
D: Annual demand Rate C: Unit Cost
Q: Batch size H: Annual holding & Storage cost/ unit ( H = I *C )
S: Set-up Cost i: Percent carrying cost (interest rate)

• Cycle Stocks: Tradeoff b/w Fixed Cost and Holding Costs

ex) Inventory Profile of Retail shop
D (annual D Rate): $1,200/year
S (Set-up Cost): $2,000/order
C (Unit cost): $200/unit
I ( % Carrying Cost): 25%

• Cost minimization Goal “ Smallest Possible TC”

• The Newsvendor Model

Lead Time and Demand Uncertainty

• Setting:

• Decisions: Order Quantity & Reorder Point (ROP)

We want to have enough inventory in our system to cover demand during lead time (DLT)

• Cycle Service Level: Probability of not stocking out in a cycle

• Given cycle service level (CSL), Choose ROP such that

• Average Inventory:
- Pipeline Inventory: amount of inventory on transit
- Cycle Stock: amount of inventory that fluctuates over time because of the ordering time
- Safety Stock: “Just In Case” Inventory

• Average pipeline Inventory = Demand Rate * Flow Time

- Suppose we can identify “Order Quantity in Transit” and pay for the inventory when it is shipped,
then we should also include cost of pipeline Inventory.
- Inventory on route!
 VII. Supply Chain Management
Supply Change Management

• The system of suppliers, manufacturers, transportation, distributors, and vendors that exist to
transport RM to final products and supply those products to customers
• The Information, money and people associated with the system.
(Value chain, the logistics network, the distribution network)

Supply Chain
• Consists of:

• Aims to matching S & D

• Achieves the RIGHT product + Price + Place + Quantity + Customer + Time = Higher profit!
• Supply Chain Challenges & Conflicting Objectives

• The Bullwhip Effect: The variance of order is greater than that of sales, and the distortion increases
as one moves upstream.

• Causes & Preventions

- Forecasting updates and Lead Time Delays: Avoid misleading forecast updates
- Order Batching (FC): Reduce / Eliminate FC
- Price Fluctuations (Forward buying): Stabilize prices
- Rationing & Shortage Gaming: Eliminate gaming & Shortage situations
 VIII. Quality Management

Historical View Modern View

When something goes wrong, When something goes wrong, no one is to blame.
it is the fault of the individual The system as a whole is responsible
Production must not be interrupted for any Shut down production if necessary. Find and fix
reason. Move off problems from the production the root so the problem does not recur
Heroes are who resolve crises Heroes are who prevent crises
Use inspection to remove defects Prevent defects by control the process
Test and rework a complex process until it is right Do it right the first time; design a process that
avoids mistakes
Perfect quality would be too expensive to achieve Perfect quality is a goal as it forces to learn
continuously
Find the right way to do something and keep Production systems should evolve continuously.
doing it that way Improving quality leads to learning.
Dedicated quality department is responsible for Achieving quality is an integral part of operations
quality and a shared responsibility of everyone in the
organization
Problem solving is an art Problem solving is more a science than art. It
must be based on objective data
 Sample
SampleFinal Questions
Midterm Questions

1. Production process of manufacture company

(a) What is the maximum throughput rate for product?

(b) What is the cycle time of this process?
(c) What is the theoretical flow time?
(d) Assume the operator of Machine C has come up with an innovation that has reduced processing
time from 7.5 minutes to 5 minutes.

As a result, the flow time will ___________, but throughput rate will ___________.

(e) When the demand rate = capacity rate of the process, what is the short run utilization of resources?
2. Consider the following flow diagram for processing a customer order, where each order is either Type
A or Type B.

The Exact Processing times and the resources needed for each activity are listed in the table below

There is one unit of resource category I in the system and there are two units of resource category II in
the system. Customers currently arrive at the rate of 2.5 per hour; one-fifth of the customers are Type A,
and four-fifths of the customers are Type B.

(a) what is the bottleneck resource?

(b) What are the implied utilization for each resource listed above?
3. Suppose the store opens at 8AM. Customers show up at the rate of 30 per hour until 1PM, and then at
the rate of 45 per hour until 3PM. The store closes at 3PM regardless of the number of customers
waiting in line, and the unsatisfied customers are sent away. Suppose that every customer who shows
up at the store joins the line and waits until satisfied or sent away. The store can serve customers at the
rate of 50 per hour between 8AM and 9AM, at the rate of 10 per hour between 9AM and 12 Noon, and
then at the rate of 40 per hour between 12 Noon and 3PM. Use the continuous-time model in your
calculation.

(a) How many customers do you expect to see in the line at 11:30 AM? How many customers are sent
away at the end of the day?

(b) Calculate the average number of waiting customers between 8AM and 3PM. Then compute the
average amount of time a customer spends on the line.
4. The bank is consider opening a drive-through window for customer service. Management estimates
that customers will arrive at the rate of 15/hour. The teller who will be staffing the window can serve
customers at the rate of one every three minutes. Assuming Poisson arrivals and exponential service,
find:

(a) Utilization of the teller

(b) Average number in the waiting line

(c) Average number in the system

(d) Average waiting time in line

(e) Average waiting time in the system including service

(f) Probability that the time between two successive arrivals exceeds 5 minutes

(g) Suppose now that the service times are deterministic (at the same rate). Which of the above
answers would change, and what would be the new values for these answers?
5. Your company has just received an order for a specially designed electric motor. The contract states
that if the project is not completed 12 days from now, then beginning on the 13th day, your firm will
experience a penalty of $300 per day. The table below shows the activities that need to be completed
for this project, along with data on direct costs and precedence relationships.

Activity Normal Normal Crash Time Crash Immediate

Time (days) Cost ($) (days) Cost ($) Predecessors
A 4 1,000 3 1,300 None
B 7 1,400 4 2,000 None
C 5 2,000 4 2,700 None
D 6 1,200 5 1,400 A
E 3 900 2 1,100 B
F 11 2,500 6 3,750 C
G 4 800 3 1,450 D, E
H 3 300 1 500 F, G

(a) Draw the critical path diagram

(b) What completion date do you recommend for the project? What is the project cost for that project
length?
6. Division X is in constant need of trained employees. The demand rate for trained employees is
relatively constant at 5 each month. The division can run a training program that costs $2,500 per
offering, regardless of class size. Students graduate with a certificate and a $100/ week pay raise. This
raise is operative even if they go back to their old job, where they will work until they are reassigned to a
higher-skilled position. Assume 52 weeks per year.

(a) What class size is optimal and how many classes should be offered each year?

7. Due to the successful advertising campaign, customer demand for lemonade increased from 4 crates
to 8 crates per week (and remained at 9 crates per week during the next 30 weeks). The demand at the
hard lemonade bottling plant showed a much higher fluctuation. In other words, the variability of demand
at the bottling plant was higher than the variability of demand at the retail store.

(a) What is this effect called?

The bullwhip effect (variability of demand at the manufacturing was greater than retailer)

(b) Describe the relationship between lead times (for production and delivery) and the observed effect.

A longer LT increases the magnitude of the bullwhip effect.

(c) List key actions that could be taken in order to avoid this effect.

Open communication to improve the flow of information and reduce variability, Set the common goal,
Reduce order sizes
8. Daily demand for ice cream at I-SCREAM parlor is normally distributed with a mean of 100 quarts and
a standard deviation of 120 quarts. The owner has the ice cream supplies by a wholesaler who charges
$2/quart. The wholesaler charges a $100 delivery charge independent of order size. The opportunity cost
of capital to I-SCREAM is estimated to be 25% per year. Assume 360 days in the year.

(a) What is the optimal order size of each order?

(b) Assume that it takes 9 days for an order to be supplied. The owners would like to ensure no stock-
outs in 95% of the lead times. What is the optimal safety stock the store should have?

(c) Currently the owners do not follow the EOQ and safety stock policy outlined above. They order 2,000
quarts of ice cream when they have 1,000 quarts on hand. Assume that it takes 9 days for an order to be
supplied. What is the average time spent by a quart of ice cream at the parlor?
9. The Luxury hotel has a policy of overbooking the room rentals to make up for customers who would
make reservations but fail to show up. The average lost contribution for a vacant rom was $20 per night if
a customer reserved a room but was a “no-show”. About 10% of the quests who did show up (with
reservation) but could not be honored because the hotel was full, could be placated without any extra cost
to Gates. Another 30% were satisfied with being transferred to another hotel at a cost to the Gates of $3
per reservation. The remaining 60% of the quests were so upset that the hotel could expect a loss of
future business with a NPV of $50. The manager reviewed his records and found the following distribution
of no-shows.

Number of No-shows Probability Cumulative Pr

1 0.2 0.2
2 0.2 0.4
3 0.3 0.7
4 0.2 0.9
5 0.1 1.0

(a) What is the expected loss in revenue by not accommodating a customer with a reservation?

(b) What is the optimal level of overbooking that you would recommend to the hotel?

(c) As a hotel manager, you realize that $50 may be an underestimate of loss of future business for upset
customers. You impute a cost of loss of good will of $200. Considering this, your optimal level of
overbooking as computed above would ( Increase/ Decrease / Remain unchanged )
10. Galen has been hired by CMP airlines to redesign their terminal counter operations, where
customers check in. The airline is very sensitive to their customer’s time spent waiting for service and
being served as many customers dislike waiting in the queues. CMP Airline currently has three counters,
each staffed by a single employee, and a single FCFS queue for all customers to join. Galen redesigns
the queues so that there is a dedicated queue and a dedicated server for each of the customer classes
(economy/business/first class). Customers cannot cross between servers.

Economy class passengers arrive at a rate of 19/hour, Business class passengers arrive at a rate of
13/hour, and First-class passengers arrive at a rate of 4/hour. Each of the three serving employees takes
an average of 3 minutes to check in customers. Assume arrivals follow a Poisson process and service
time is exponentially distributed. Compare the total time (waiting and service) customers spend in the two
system designs. RISK POOLING!

😊 GOOD LUCK to you ALL! 😊