SDM 5001 SYSTEMS ARCHITECTURE
LECTURE 9.1
PRODUCTION PROCESS ANALYSIS
© LGChan
LITTLE’S LAW
2
© LGChan
Little’s Law
Little's Law is a theorem from queuing theory which states that
the average number of customers L in a system (over some interval) is equal to
their average arrival rate R , multiplied by their average time in the system T
John D.C. Little
I = Inventory (MIT) 1954, 1961
I=RxT R = Throughput rate, or Flow Rate
T = Flow Time, time it takes a flow unit to complete a process
note: some texts use T as “lead time” or “cycle time” which is incorrect
Applications
R or TH = throughput (arrival rate)
TH is the velocity or speed of production and is calculated by determining how many items are produced and dividing it by the
length of time it took to produce them
T or CT = cycle time (average time in the system)
CT is the time it takes to complete the production cycle or the average time it takes to produce one unit
I or WIP = work in process (average number of units or customers in a system)
WIP is the number of items currently in production or being serviced in some way
Throughput TH = WIP/CT
Cycle Time CT = WIP/TH
WIP WIP = CT x TH 3
© LGChan
Example 1 : Little’s Law and Productivity
Example 1 Normal Demand
Assume, Throughput (TH) produce 50 units per day, and Work in Process (WIP) remains relatively constant at 200 units
Cycle time (average time it takes to complete one unit) = 200/50 = 4 days
This means we can accommodate new orders of 50 units each day and the system will remain balance
Example 2 New Demand
Suppose we receive 2 (two) orders for 85 units, 35 units per day more than the standard order of 50 units per day
WIP would increase from 200 units to 270, and because TH would remain constant at 50 units per day,
CT would immediately increase from 4 days per unit to (270/50) = 5.4 days per unit
This means efficiency decrease
Lesson
Significant levels of new orders cause production efficiency to decrease
(unless we decrease work in progress WIP or increase throughput with more productivity )
4
© LGChan
Example 2: Little’s Law and Bottleneck
Design Dye Fabric Cutting Sewing Shirt
5 min/shirt 4 min/shirt 6 min/shirt
Dye Fabric Cutting Sewing System Capacity
Throughput 5 min/shirt 4 min/shirt 6 min/shirt
Processing Capacity (R) 12 shirts/hr 15 shirts/hr 10 shirts/hr 10 shirts/hr
Average Processing Time, T = 5 + 4 + 6 = 15 minute per shirt = 15/60 = 0.25 shirt/hr
Work in Progress in the System using Little Law, WIP or I = R x T = 10 x 0.25 = 2.5 shirts/hr
This Production Process is uneven because the cycle time of the 3 processes are not synchronize
o Dye Fabric and Cutting machines are not fully utilized in the production
o Work in Progress accumulate at Sewing, when Dye Fabric and Cutting are fully utilized
The bottleneck is Sewing process
5
© LGChan
KINGMAN EQUATION
6
© LGChan
Application of Kingman Equation in Lean Production
Kingman Equation, from mathematical queuing theory, provides an
approximation of the waiting time of the parts for a single process based on its
utilization and variance
John Kingman, 1961
p = average processing time for a resource
u = Resource Utilization
= Flow Rate/ Resource Capacity
(ie 100%>resource capacity > demand rate)
v = Variability Factor
variation in arrival of units
= variation in processing times
Variation in arrival units = 1 + [change in arrive/average arrival]
Variation in processing time = 1 + [change in processing time/average processing time]
7
© LGChan
Production System 1 – Maximum Production Capacity
Design Dye Fabric Cutting Sewing Shirt
5 min/shirt 4 min/shirt 6 min/shirt
Resource Dye Fabric Cut Fabric Sewing
Processing Time (min/shirt), p 5 4 6
Capacity (shirts/min) 1/5 1/4 1/6
Capacity (shirts/hr), v 12 15 10 bottleneck
Resource Utilization, u 10/12=83% 10/15=67% 10/10=100%
8
© LGChan
Production System 2 – Decrease Customer Demand
Assume Customer Demand is 8 shirts/hr (below bottleneck production 10 shirts/hr )
Flow Rate is the minimum of process capacity, input rate, or demand rate = 8 shirts/hr
Resource Utilization at Sewing is u = 8/10 = 0.8 (a result of restricted input)
Input Fabric is supplied at a rate of 8 shirts/hr:
Variation in input rate = 1 - 8/10 = 0.2 (20% decrease)
Variation in processing time = 1 (no changes)
Variability Factor, v = (variation in input rate) / (variation in processing times)
= (0.2) / (1+0) = 0.2
Resource Dye Fabric Cut Fabric Sewing
Processing Time (min/shirt), p 5 4 6
Capacity (shirts/min) 1/5 1/4 1/6
Capacity (shirts/hr), v 12 15 10 bottleneck
Resource Utilization, u 8/12=67% 8/15=53% 8/10=80%
Production Delay Time of Factory or Waiting Time for Customer
= 6 x (8/10) / [1 – 8/10] x 0.2
= 6 x 4 x 0.2
= 4.8 minutes per shirt 9
© LGChan
Utilization with New Demand
o Given that there is enough capacity on average to
meet demand (capacity > demand rate)
o With variability, there is an exponential increase in the
queue time (Twait) with increasing capacity utilization
o Roughly at 80% utilization the queues become critical
o When Utilization → 100%, Twait approaches infinity
How to Reduce Waiting Time
o Reducing process time, p, through process improvement and
elimination of waste will reduce queuing
o Keep utilization as low as possible, to keep u/(1-u) as low as possible
o Reduce arrival variation (e.g. scheduling slots/appointments) and
processing time variation (eg. Standard Work)
10
© LGChan
Application of Kingman Equation for Lean
Eliminate Muri (overburden) and Mura (unevenness) to avoid long and variable lead
times and the poor and unreliable service
1. Plan at less than 100% utilization
2. Sell only what can be sold
3. Plan using realistic cycle times
4. Reduce utilization by continuously reducing waste
5. Level sales and level load (through Heijunka and Kanban)
6. Standardize, and continuously detect and eradicate problems to reduce variation
7. Adapt your utilization to the level of variation you experience
11
© LGChan
Online Kingman Equation Simulation of Traffic Flow
In this simulation, you can play with utilization (particularly the “density”
parameter in the model) and variation (using the mix between passenger cars
and trucks using “truck fraction”) and evaluate the effects on road congestion
http://www.traffic-simulation.de/
12
© LGChan
END OF LECTURE 9.1
PRODUCTION PROCESS ANALYSIS
13
© LGChan
SDM 5001 SYSTEMS ARCHITECTURE
LECTURE 9.2
THEORY OF CONSTRAINTS
© LGChan
Theory of Constraints: A Process of Ongoing Improvement
Goldratt, Eliyahu, J Cox (1993) The Goal: Kim, Gene, K Behr, G Spafford (2014) The
A Process of Ongoing Improvement. Phoenix Project: A Novel about IT, DevOps, and
Aldershot:Gower, c2004. Helping Your Business Win. Portland, Oregon:IT
PR9510.9 G621G Revolution Press. HD30.2 Kim 2014
2
© LGChan
Introducing Theory of Constraints
Theory of Constraints is a management concept to identify constraints and alleviate
them to the extent possible so as to achieve maximum throughput from the system
Video (3:01) https://www.youtube.com/watch?v=Cb2rg4YWfCo
3
© LGChan
Theory of
Constraints
Process
What to What to Change How to Cause
Change? To? the Change?
From observing and By examining the An action plan is
analysis, cause-and- logical assumptions developed to transition
effect is used to behind the conflicts, the system from where it
identify the underlying counter-measures to is today to realize the
common cause or root the root causes are goal which is aligned to
causes. identified. remove the constraints
This often involves
changes to design,
measurements and
structures
4
© LGChan
Five Step Process in Theory of Constraints
1 Identify the system's constraints 2 Decide how to exploit the
o What is the Goal? system's constraints
o What is Throughput? o What is the constraint?
o What is Inventory? o How do we get as much throughput
o What is Operating Expense? as possible?
5 If a constraint is broken in Step 3 Subordinate everything else to
4, go back to Step 1 the decisions of Step 2
o What might happen if the o Throughput?
constraint is elevated? o Inventory?
o Operating Expense?
4 Elevate the system's constraints
o Throughput?
o Inventory?
o Operating Expense?
5
© LGChan
DRUM ROPE BUFFER
6
© LGChan
Drum Rope Buffer
1. Every system has a bottleneck or constraint
2. A bottleneck is a state of affairs where demand for service exceeds the capacity to serve
3. The Throughput of a system is dependent on the Throughput of the Bottleneck
4. Given (1), (2), & (3), for maximum output, a system ought to keep the bottleneck working at 100%
capacity with little or no defects (scrap, waste (muda), time-traps)
5. Given (4), Non-bottleneck processes should be working at less than 100% capacity, so as to not over-
burden the bottleneck with large batches of work-in-process (WIP)
o Synchronizes by beating a drum keeping pace with the slowest soldiers
o Prevents the troop from spreading by tying other soldiers with the
slowest solders (ie bottleneck) by a rope
o Length of rope provides a buffer and constraint to the range of
fluctuation in the speed of each soldier
7
© LGChan
Theory of Constraints is based on Concept of Drum, Buffer and Ropes
Drum Buffer Ropes
When upstream processes cannot The Rope is a method by which the
The Bottleneck or Constraint, acts as
produce as much as is needed by the Constraint can signal to the upstream
a Drum : it sets the rhythm that the processes (non-bottleneck processes)
whole system should follow Bottleneck
The Constraint is starved and overall when to slow down, when to stop, or
Output of the constraint is the when to produce faster and the quantity
system output is compromised
drumbeat We need a buffer of inventory/work in
o Sets the tempo for other operations process in front of constraint to take care
o Tells upstream operations what to of variation is demand
produce
o Tells downstream operations what to
expect
Rope is called Pull Scheduling in Lean
In Lean Manufacturing, this is also called Buffer inventory will help to level-out
Manufacturing terms or a Stack in
Takt Time variation and is related to Lead Time
software programming
8
© LGChan
Constraints
A constraint can be a thing, process, person, structure, object that prevents the flow of value to the customer
Any system can produce only as much as its critically constrained resource
Maximum speed of the process is the speed of the slowest operation
Any improvements will be wasted unless the bottleneck is relieved
Internal Constraints External Constraints
Process constraints: Machine time, maintenance Material Constraints: Insufficient supply
Policy constraints: No overtime, no budget Market Constraints: Insufficient demand
Constraint
A Pull System
60 units 70 units 40 units 60 units
per day per day per day per day
Maximum Throughput = 40 units per day
9
© LGChan
Example : Shop Floor Control
Drum-Buffer-Rope for Shop Floor Control
o Drum: Pace Setting Resource - constraint
o Buffer: Amount of protection in front of the resource
o Rope: Scheduled staggered release of material to be in line with the Drum’s schedule
A Pull System
Buffer
60/day 70/day 40/day 60/day
Rope Constraint
(Drum)
10
© LGChan
How to Manage Bottlenecks
The four steps in managing bottlenecks are:
1. Recognize that the bottleneck operation determines throughput contribution of the system as a whole
2. Search and find the bottleneck operation by identifying operations with large quantities of inventory
waiting to be worked on
3. Keep the bottleneck busy and subordinate all non-bottleneck operations to the bottleneck operations
4. Take actions to increase bottleneck efficiency and capacity – the objective is to increase throughput
contribution minus the incremental costs of taking such actions
Drum Buffer Rope Study* shows
o Reduce Lead times an average of 70%
o Reduce inventory an average of 49%
o Reduce Operating Expenses
o Improve on time delivery performance an average of 44%
o Increase Revenue an average of 63%
* Mabin & Balderstone. The World of the Theory of Constraints. St. Lucie Press 2000
11
© LGChan
Production in Traditional Fashion Industry vs Zara
Lean Manufacturing in Zara
12
© LGChan
MEASURING PERFORMANCE IN THEORY OF CONSTRAINTS
13
© LGChan
Measuring Performance in Theory of Constraints
1 Throughput
The rate at which the system generates money through sales (value added)
Only Money generated by the system and into the organization are included
Throughput = Sales Revenue – Cost of Goods Sold
Not Throughput:
Investment in Machinery, Finished Goods in Warehouse not sold, Work in Process, Raw Materials
2 Inventory (sometimes known as Investment)
All the money the system has invested in purchasing things which it intends to sell
In Theory of Constraints, Inventory is a “liability” (not an asset) because it is not producing revenue until it is
sold
Raw materials, work in process, finished goods and scrap are “Inventory”
3 Operating Expenses
All the money the system spends in order to turn inventory into throughput
All employee time is “OE” (through salaries and wages: direct, indirect, operating, etc.)
Depreciation of a machine is “Operating Expenses”
Operating supplies are “Operating Expenses”
14
© LGChan
Financial Implications of Improvements from Theory of Constraints
Improvements Changes and Impacts
Throughput Inventory Op Expense Net Profit Return on Inventory Cash Flow
Increase No change No change Increase Increase Increase
No change No change Decrease Increase Increase Increase
No change Decrease* No change No change Increase Increase
* Note : if the carrying cost of inventory decreases, then net profit will increase
Video (5:22) https://www.youtube.com/watch?v=07EgECNPd8k
Rank the Order of Importance of Improvements in Theory of Constraints?
1. Throughput
2. Inventory or Assets
3. Operating Expenses
Why Do We Need Inventory To Protect Throughput?
We need inventory to protect throughput, because there are statistical fluctuations
(uncertainties) and dependent resources (cellular manufacturing)
15
© LGChan
Return on Investment
Throughput – Operating Expenses
Return on Investments =
Investments
What is the Priority to Increase Return on Investment?
1 Decrease Operating Expenses
Traditional Management 2 Decrease Investments
3 Increase Throughput
1 Decrease Investments
Japanese Management 2 Increase Throughput
3 Decrease Operating Expenses
1 Increase Throughput
Constraint Management 2 Decrease Investments
3 Decrease Operating Expenses
16
© LGChan
Performance Management: How to Improve Throughput Value
Throughput Value Sales − Direct Material Costs
per factory hour = Usage of bottleneck in hours (factory hours)
This enables businesses to take short-term decisions when a resource is in scarce supply
Factory hours are measured in terms of use of the bottleneck resource.
Throughput Value per factory hour
Throughput Ratio (TAR) = Cost per factory hour
The higher the ratio, the more profitable the company.
If a product has a ratio of less than one, you lose money every time a product is made
How to Increase Throughput Value
o Increase selling price of the product
o Reduce material cost per unit
o Reduce factory cost
o Improve productivity at the bottleneck 17
© LGChan
EASY EXAMPLE : THROUGHPUT ACCOUNTING
ONE PRODUCTION LINE WITH ONE MACHINE
APPLICATION OF THEORY OF CONSTRAINTS
18
© LGChan
Example : Maximizing Throughput
Throughput accounting is an approach to accounting which is largely reflects with
Just In Time philosophy
Product A Product B Product C
Selling Price per unit $30 $28 $27
Raw Materials cost per unit $6 $8 $6
Processing Capacity per unit 2 4 3
Average Sales Demand units 13000 10000 8000
Processing time from machine is a bottleneck resource
Processing hours available is 80,000
Operating Expenses is $480,000
What is the Optimal Production Plan and Maximum Profit?
19
Video (5:01) https://www.youtube.com/watch?v=Mc36-1PtoPA
© LGChan
Throughput Calculations and Ranking
1 Calculate Throughput Value per unit
Product A Product B Product C
Throughput Value per unit $30 – 6 = $24 $28 – 8 = $20 $27 – 6 = $21
2 Calculate Throughput Value per hour
Product A Product B Product C
Throughput Value per hr $24/2=12 hr $20/4 = 5 hr $21/3 = 7 hr
3 Rank the Product by Their Throughput Value per unit
Product A Product B Product C
Throughput Value per unit 1st 3rd 2nd
20
© LGChan
Machine Capacity Allocation and Production Plan
First we produce 13,000 units Product A: 13,000 x 2 = 26,000 hours
Second we produce 8,000 units Product C: 8,000 x 3 = 24,000 hours
Total Processing Hours of Product A and Product C: 26,000 +24,000 = 50,000 hours
Remainder Processing Capacity 80,000 – 50,000 = 30,000 hours
Finally we produce Product B from remaining 30,000 hours = 30,000 / 4 = 7,500 units
21
© LGChan
Calculate Profit from Throughput Value
Product A Product B Product C
Total Throughput Value
per product 13,000 x $24 = $312,000 7,500 x $20 = $150,000 8,000 x $21 = $168,000
Total Throughout $312,000 + $150,000 + $168,000 = $630,000
Operating Expenses = $480,000
Profit = Total Throughput value – Operating Expenses = $630,000 - $480,000 = $150,000
Throughput Ratio = Throughput Value per Hr / Cost per Factory Hr
Cost per Factory hr = Total Operating Expenses / Processing Hours Available
= $480,000 / 80,000 = $6 per hr
Product A Product B Product C
Throughput Ratio 12/6 = 2 5/6 = 0.83 7/8 = 1.17
Production is Profitable if Throughput Ratio >1 Production is Unprofitable if Throughput Ratio >1
Product A and Product C are profitable, but Product B is loss making 22
© LGChan
TEST YOURSELF
23
© LGChan
A factory that makes 2 products, A and B. The products each go through two different steps.
Product A goes to Machine 1 (M1) first and then Machine 3 (M3). Product B goes to Machine 2 (M2) first and then on to M3.
Table below shows the price the items are sold for, how much raw materials
go into each, how much labor goes into each and the profit of each product.
The last column shows some monthly market constraints.
No matter what, you can only sell 140 of A or B per month.
Also, for A, you have a minimum contractual requirement to make at least
75 per month. Assume 336 hours of time per month.
What is your production schedule to
produce the maximum profit?
Bottle neck is Product A
o Product A profit per hour =
o Product B profit per hour =
o Satisfy minimum requirement of product A requires ? hours of M3
o Remaining production time at M3 for Product B = hours
o Number of B products produced at remaining time = units
24
© LGChan
END OF LECTURE 9.2
THEORY OF CONSTRAINTS
25
© LGChan
