Push and Pull Production Systems

1

• Push and Pull Production Systems

2
Overall View of MRP Program
Aggregate Planning
MPS
Bill of Material
Inventory Record
MRP Program
Planned-Order Releases
3
One Card Kanban System
Outbound stockpoint
Outbound stockpoint
Completed parts with cards enter outbound
stockpoint.
Production cards
When stock is removed, place production card in
hold box.
Production card authorizes start of work.
4
Key Difference Between Push and Pull

• Push Systems schedule work releases based on
  demand.
• inherently make-to-order
• Pull Systems authorize work releases based on
  system status.
• inherently make-to-stock

5
Push vs. Pull Mechanics
PUSH
PULL
Schedule orders, forecasts, arrivals, or other
upstream information
Status of process or other downstream stations
Process
Process
Job
Job
6
Push and Pull Examples

• Are the following systems essentially push or
  essentially pull?
• Pure MRP system
• Soda vending machine
• Photocopy shop
• Supermarket (goods on shelves)
• Runway at airport during peak periods
• Order entry server at Amazon.com
• Doctors office

PUSH
PULL
PUSH
PULL
PULL
PUSH
PUSH into office, PULL into exam room
7
Push-Pull Interface

• Concept
• Push and pull can be used in same system.
• Dividing point is called the push-pull interface.
• Benefit choosing the location of PP interface
  wisely can enable a system to take strategic
  advantage of the benefits of pull, while still
  retaining the customer-driven character of push.

8
Example Quick Taco Production Line
PP Interface
Pull
Push
Cooking
Packaging
Assembly
Sales
Refrigerator
Warming Table
Customer
Inventory buffer
Replenishment signal
Material flow
Workstation
9
Example Custom Taco Production Line
PP Interface
Pull
Push
Cooking
Packaging
Assembly
Sales
Refrigerator
Customer

• Notes
• PP interface can differ by time of day (or
  season).
• PP interface can differ by product.

10
Example HP Printer Supply Chain
U.S. DC
Customer
Printed CircuitAssembly Test
Final Assembly and Test
Integrated CircuitManufacturing
European DC
Customer
Far East DC
Customer
PP Interface

• Notes
• PP interface located in markets to achieve quick
  response to customers
• Delayed differentiation of products (power
  supplies for different countries) enables pooling
  of safety stocks

11
Push-Pull Interface Conclusions

• Basic Tradeoff
• responsiveness vs. inventory (time vs. money)
• moving PP interface closer to customer increases
  responsiveness and (usually) inventory
• Optimal Position of Push-Pull Interface
• need for responsiveness
• cost of carrying inventory ? product
  diversification
• Levers
• product design (postponement)
• process design (quick response manufacturing)

12
Japanese Success

• Low Unit Cost
• low inventory
• reduced space
• little rework
• High External Quality
• high internal quality
• pressure for good quality
• promotion of good quality (e.g., defect detection)

• Good Customer Service
• short cycle times
• steady, predictable output stream
• Flexibility
• avoids committing jobs too early
• encourages floating capacity

13
The Magic of Pull

• Pulling Everywhere?
• You dont never make nothin and send it no
  place. Somebody has to come get it.
• Hall 1983

• No! Its the WIP Cap
• Kanban WIP cannot exceed number of cards
• WIP explosions are impossible

14
Pull Benefits Achieved by WIP Cap

• Reducing Manufacturing Costs
• prevents WIP explosions
• reduces average WIP
• reduces engineering changes
• Reducing Variability
• reduces cycle time variability
• pressure to reduce sources of process variability
  
• promotes shorter lead times and better on-time
  performance

15
Pull Benefits Achieved by WIP Cap

• Improves Quality
• pressure for higher quality
• improved defect detection
• improved communication
• Maintains Flexibility
• avoids early release
• less direct congestion
• less reliance on forecasts

16
CONWIP

• CONWIP (constant work in process) allow next job
  to enter line each time a job leaves (i.e.,
  maintain a WIP level of m jobs in the line at all
  times).
• Assumptions
• 1. Single routing
• 2. WIP measured in units

17
CONWIP Controller
Work Backlog
PN Quant



PC
PC
. . .
Workstations
18
Push and Pull Schematics
CONWIP
. . .
Pure Push (MRP)
. . .
Pure Pull (Kanban)
. . .

Full Containers
Authorization Signals
19
Mean-Value Analysis Model

• Mean-Value Analysis (MVA) an iterative procedure
  that develops the measures of the line with WIP
  level w in terms of those for WIP level w-1.
• Basic Approach Compute performance measures for
  increasing w assuming job arriving to line sees
  other jobs distributed according to average
  behavior with w-1 jobs.

20
Notation
21
Mean-Value Analysis Formulas
22
Iterative Procedure

• Let with WIPj(0) 0 and TH(0) 0. Start with w
  1.

Next,

23
Using MVA to Evaluate Line Performance
Best case
Practical worst case (ce(j) 1)
Reduced variability (ce(j) 0.5)
Worst case
24
Comparisons of CONWIP with MRP

• A fundamental distinction between push and pull
  systems
• Push systems control throughput and observe WIP.
• Pull systems control WIP and observe throughput.
• Observability
• It is preferable and control the robust parameter
  and observe the sensitive parameter.
• WIP is robust and observable, while throughput is
  sensitive.

25
CONWIP Efficiency Example

• Equipment Data
• 5 machines in tandem, all with capacity of one
  part/hr (uTHteTH)
• exponential (moderate variability) process times
• CONWIP System
• Pure Push System

PWC formula
5 M/M/1 queues
26
CONWIP Efficiency Example (cont.)

• How much WIP is required for push to match TH
  attained by CONWIP system with WIPw?

• In this example, WIP is always 25 higher for
  same TH in push than in CONWIP

27
CONWIP Efficiency

• Law (CONWIP Efficiency) for a given level of
  throughput, a push system will have more WIP on
  average than an equivalent CONWIP system.
• Corollary For a given level of throughput, a
  push system will have longer average cycle times
  than an equivalent CONWIP system.

28
CONWIP Variability

• In a CONWIP system
• Negative correlation between WIP levels at
  different stations.
• Dampen fluctuations in cycle time.
• In a push system
• WIP levels at individual stations are independent
  of one another.
• Cycle times are more variable.
• Increased cycle time variability results in
  longer lead times.

29
Cycle Time and Lead Time
90 customer service
CT 10 ?CT 3
Lead time 14 days
Lead time 23 days
CT 10 ?CT 6
30
CONWIP Robustness

• Law (CONWIP Robustness) A CONWIP system is more
  robust to errors in WIP level than a pure push
  system is to errors in release rate.
• Increased robustness is probably the most
  compelling reason to use a pull system, such as
  CONWIP, instead of a push system.

31
CONWIP Robustness
p marginal profit per job h cost for each unit
of WIP

• Profit Function
• CONWIP
• Push
• Key Question what happens when we dont choose
  optimum values (as we never will)?

need to find optimal WIP level
need to find optimal TH level (i.e.,
release rate)
32
CONWIP vs. Pure Push Comparisons
Optimum
CONWIP
Efficiency
Robustness
Push
33
Comparisons of CONWIP with Kanban

• CONWIP and Kanban are both pull systems.
• Compare to a pure push system, both
• achieve a target throughput level with less WIP.
• exhibit less cycle time.
• are easier to manage as WIP is a more robust
  control.
• Differences
• Card count issues
• Product mix issues
• People issues

34
Card Count Issues

• Parameter settings
• In a one-card kanban system, one must establish a
  card count for every station.
• In a CONWIP system, there is only a single card
  count.
• Types of cards
• Cards are typically part number-specific in a
  kanban system.
• Cards are line-specific in a CONWIP system.

35
CONWIP Controller
Work Backlog
PN Quant



PC
PC
. . .
Workstations
36
Product Mix Issues

• Kanban is applicable only in repetitive
  manufacturing environments (steady material
  flows, fixed paths).
• We generally want to put more production cards
  before and after the bottleneck station, in order
  to protect it against starvation and blocking.
• But which station is the bottleneck?

37
Product Mix Example
Processing times Product A
1
3
2.5
1
1
Processing times Product B
1
1
2.5
3
1

• A 50-50 mix of products A and B

Average time on machine 2 0.5(3) 0.5(1)
2 Average time on machine 3 0.5(2.5) 0.5(2.5)
2.5 Average time on machine 4 0.5(1)
0.5(3) 2
Machine 3 is the bottleneck.
38
Product Mix Example (cont.)

• Which station is bottleneck depends on product
  mix.

? The card counts may vary over time in a kanban
system.

• CONWIP has only a single card count that depends
  on production rate.

? The CONWIP system is simpler to manage than a
kanban system.
39
People Issues

• The fact that kanban systems pull at every
  station introduces a certain amount of stress
  into the system.
• A CONWIP system acts as a push system at every
  station except the first one, thus is subject to
  less pacing stress.
• Closer relationship between operators of adjacent
  workstations in a kanban line can be achieved by
  other learning motivators.

40
Workstations Connected by a Finite Buffer
Outbound stockpoint
Completed parts with cards enter outbound
stockpoint.
Machine 1
Machine 2
Production card authorizes start of work.
When stock is removed, place production card in
hold box.
Production cards for Machine 1
Production cards for Machine 2