Вы находитесь на странице: 1из 29

Measures and Basic Dynamics

of Operation Systems

Dr. Jeff Hong


IELM Dept, HKUST
Fall 2010

Note01: Dynamics 1
VIA Example
„ Vancouver International Airport
‰ The airport is one of the top 10 airports in the world
‰ The airport authority wants to maintain its excellent
customer service standards
‰ Customers complain that the waiting time in the airport
security checkpoints is too long
‰ The airport management team wants to reduce the
waiting time

Note01: Dynamics 2
VIA Example
„ To analyze the problem, we need to answer
three important questions:
‰ On average, how many passengers pass through the
checkpoints per unit of time?
‰ On average, how much time does a passenger spend
within the boundaries of checkpoints?
‰ On average, how many passengers are within the
boundaries of checkpoints at any point in time?

Note01: Dynamics 3
Three Key Performance Measures
„ Throughput (TH)
‰ The number of units that pass through the system per
unit of time
„ Work-in-process (WIP)
‰ The total number of units present within the
boundaries of the system
„ Cycle time (CT)
‰ The total time spent by a unit within the boundaries of
the system

Note01: Dynamics 4
VIA Example
To understand customer flow, the airport managers began by analyzing a
single security screening line, which is comprised of an X-ray scanner with an
operator and screening officers.
‰ Customers can have 0, 1, 2 or 3 carry-on bags, including purses, wallets
and so on. But on average, a typical customer has 1.5 bags.
‰ The X-ray scanner can handle 18 bags per minute.
‰ For a flight with 200 passengers, the following approximate arrival rate
pattern: about 75 passengers arrive 80-50 minutes before scheduled
departure, 100 arrive 50-30 minutes before scheduled departure, and the
remaining 25 arrive between 30-20 minutes before scheduled departure.
‰ To minimize layover time for passengers switching flights, many of VIA’s
flights depart around the same time. In this example, we will assume for
simplicity that exactly three flights, each carry 200 passengers, are
scheduled for departure each hour. (We analyze the departures at 10am)

Note01: Dynamics 5
Analysis of the VIA Example
„ Throughput (TH)
‰ The inflow rate of passengers at a security check point
changes over time.
‰ 300 passengers arrive during 9:10 – 9:30 am, giving
an inflow rate of roughly 15 passengers per minute
‰ 300 passengers arrive between 8:40 – 9:10 and
9:30 – 9:40, given an inflow rate 7.5 passengers per
minute
‰ The outflow rate, however, is limited by the X-ray
scanner with an average 12 passengers per minute

Note01: Dynamics 6
Analysis of the VIA Example
„ Work-in-process (WIP)
‰ Let I(t) and O(t) denote the inflow and outflow rate at t
‰ Let ∆(t) = I(t) – O(t) denote the difference
‰ If I(t) > O(t), then WIP is accumulated at a rate ∆(t) > 0
‰ If I(t) = O(t), then WIP remains unchanged
‰ If I(t) < O(t), then WIP is depleted at a rate ∆(t) < 0

Note01: Dynamics 7
Analysis of the VIA Example
WIP
(Passengers
in queue)

60

∆ = 3/min ∆ = – 4.5/min

∆=0/min ∆=0/min
t
8:40 9:10 9:30 9:43:20 10:00

8:40 – 9:10 : I(t) = O(t) = 7.5/min, TH(t) = O(t) = 7.5/min


9:10 – 9:30 : I(t) = 15/min, O(t) = 12/min, TH(t) = O(t) =12/min
9:30 – 9:43:20 : I(t) = 7.5/min, O(t) = 12 min, TH(t) = O(t) =12/min
9:43:20 – 10:00: I(t) = O(t) = 7.5/min, TH(t) = O(t) = 7.5/min

Note01: Dynamics 8
Analysis of the VIA Example
„ Cycle time (CT)
‰ We look at the queueing time for each customer
‰ When queue is empty, then queueing time = 0
‰ When queue is not empty at the time of the new arrival,
then the queueing time of a new arrival is (# in
queue/12) minute
‰ When a customer arrives at 9:30, he finds 60
customers before him. Then his waiting time is 5
minutes.

Note01: Dynamics 9
Stable System
„ Definition
‰ A stable system is a system in which, in the long run,
the average inflow rate is the same as the average
outflow rate.
„ Implications
‰ The average TH is the same as the average inflow
and outflow rates
‰ The WIP will grow without limit if the system is not
stable

Note01: Dynamics 10
Analysis of VIA Example
„ To analyze the problem, we need to answer
three important questions:
‰ On average, how many passengers pass through the
checkpoints per unit of time? (average TH)
‰ On average, how much time dos a passenger spend
within the boundaries of checkpoints? (average CT)
‰ On average, how many passengers are within the
boundaries of checkpoints at any point in time?
(average WIP)

Note01: Dynamics 11
Analysis of VIA Example
„ Average TH (th) :
‰ th = 600 passenger / hour = 10 passengers / min

„ Average WIP (wip):


‰ wip = (33 min x 30 passengers + 27 min x 0 passengers) / 60 min
= 16.5 passengers
„ Average CT (ct) : ct = ?
WIP
(Passengers
in queue)
60

∆ = 3/min ∆ = – 4.5/min

∆=0/min ∆=0/min
10:00 t
8:40 9:10 9:30 9:43:20
Note01: Dynamics 12
Little’s Law
„ For a stable system
wip = ct x th
„ Little’s law is extremely useful in determining the average
cycle time
‰ Production system
‰ Service system
„ Inventory turns or turnover ratio
‰ Defined as th/wip (the ratio of throughput to average WIP), to show
how many times the inventory is sold during a specific period.
‰ Used widely in operations, accounting and finance
‰ By Little’s law, it equals 1/ct

Note01: Dynamics 13
Applications of Little’s Law
„ VIA Example
‰ th = 10 passengers/min
‰ wip = 16.5 passengers
‰ ct = wip / th = 1.65 minutes
‰ On average, every passenger has to wait for 1.65
minutes

Note01: Dynamics 14
Applications of Little’s Law
„ Material flow
‰ A fast food restaurant wants to verify that it is serving
fresh meat in its hamburgers (needs to know how long
they keep meat in their refrigerator)
‰ It processes an average 5,000 kg of hamburgers per
week
‰ Typical inventory of raw meat in refrigerator is 2,500kg
‰ th = 5,000 kg/week, wip = 2,500 kg
‰ By Little’s Law, ct = wip/th = 0.5 week

Note01: Dynamics 15
Applications of Little’s Law
„ Customer flow
‰ An upscale restaurant wants to know how long
customers spent in the restaurant
‰ It serves about 200 customers per night
‰ A typical night is about 5 hours
‰ On average, there are 100 customers in the restaurant
‰ th = 200/5 = 40 customers/hour
‰ wip = 100 customers
‰ ct = wip/th = 100/40 = 2.5 hours

Note01: Dynamics 16
Applications of Little’s Law
„ Job flow
‰ A branch office of an insurance company processes
10,000 claims per year. Average processing time is
three weeks. We want to know how many claims are
being processed at any given point.
‰ Assume that the office works 50 weeks per year.
‰ th = 10,000/50 = 200 claims/week, ct = 3 weeks
‰ wip = th x ct =600 claims

Note01: Dynamics 17
Applications of Little’s Law
„ Cash flow
‰ A steel company wants to know how long the money
is tied up in working capital at the factory.
‰ The factory processes $400 million of iron ore per year.
The cost of processing ore is $200 million per year.
The average inventory is $100 million.
‰ Therefore, a total of $400+$200=$600 million flows
through the system each year (th = $600 million/year)
‰ ct = wip/th = 100/600 = 2 months
‰ Inventory turns = 1/ct = 6 / year

Note01: Dynamics 18
VIA Example
„ To reduce the waiting time in the checkpoints,
the airport managers suggest to spread out the
departure times. How will that help?

Note01: Dynamics 19
Auto-Moto Case
Auto-Moto Financial Services provides financing to qualified buyers of new cars and
motorcycles. Having just revised its application-processing operations, Auto-Moto is now
evaluating the effect of its changes on service performance. Auto-Moto receives about
1,000 loan applications per month and makes accept/reject decisions based on an
extensive review of each application. Assume a 30-day working month.

Until last year (under what we will call “Process I”), Auto-Moto processed each
application individually. On average, 20% of all applications received approval. An
internal audit showed that, on average, Auto-Moto had about 500 applications in process
at various stages of the approval/rejection procedure. In response to customer
complaints about the time taken to process each application, Auto-Moto called in UST-
EEM Consulting (UEC) to help streamline its decision-making process. UEC quickly
identified a key problem with the current process: although most applications could be
processed fairly quickly, some – because of insufficient and/or unclear documentation –
took a disproportionate amount of time. UEC thus suggested the following changes to the
process (thereby creating what we will call “Process II”):

Note01: Dynamics 20
Auto-Moto Case (cont.)
1. Because the percentage of approved applications is fairly low, an Initial Review Team
should be set up to preprocess all applications according to strict but fairly mechanical
guidelines.
2. Each application would fall into one of three categories: A (looks excellent), B (needs
more detailed evaluation), and C (reject summarily). A and B applications would be
forwarded to different specialist subgroups.
3. Each subgroup would then evaluate the applications in its domain and make
accept/reject decisions.

Process II was implemented on an experimental basis. The company found that, on


average, 25% of all applications were As, 25% Bs, and 50% Cs. Typically, about 70% of
all As and 10% of all Bs were approved on review. Internal audit checks further revealed
that, on average, 200 applications were with the Initial Review Team undergoing
preprocessing. Just 25, however, were with the Subgroup A Team undergoing the next
stage of processing and about 150 with the Subgroup B Team.

Auto-Moto Financial Services wants to determine whether the implemented changes


have improved service performance.

Note01: Dynamics 21
Analysis of Auto-Moto Case

Note01: Dynamics 22
A Case Study: On-time Delivery
of An E-Cap Manufacturer

Note01: Dynamics 23
E-Cap Production System
„ Sales department complains that the production
department always delays and the on-time
delivery rate is below 70%.

„ Production department complains that the sales


department often gives them urgent orders that
cannot be finished on time and disrupt the
normal production.

Note01: Dynamics 24
Note01: Dynamics 25
Setting Delivery Time

„ The sales department sets delivery times


‰ Totally 7 processes, each one requires about 1 shift for
each batch (10 hours). So the total production time is
about 3.5 days.

‰ Normal orders: 10 days (including order processing,


material preparation and transportation)

‰ Urgent orders: can be as low as 5 days when material


is present.

Note01: Dynamics 26
What are the problems?

Production System

New order

Note01: Dynamics 27
Note01: Dynamics 28
Improvement
„ Estimate production time:
‰ total WIP = before process WIP + in process WIP
‰ production time = total WIP / production rate

„ Reducing production time


‰ Reducing in process WIP
‰ Increasing production capacity (to reduce before process
WIP)

Note01: Dynamics 29