from dbr to simplified-dbr-8 (2) - tocico.nettocico.net/from dbr to simplified-dbr-8 (2).pdf · the...

TOCICO 2009 Webcast

From DBR to SimplifiedFrom DBR to Simplified--DBRDBR

Eli Schragenheiml ki @ t i i t [email protected]

1© 2009 TOCICO. All rights reserved.

OutlineOutlineTOCICO 2009 Webcast

• A historical perspective• A historical perspective.

• What to change?

What to change to? The toolkit of S DBR• What to change to? The toolkit of S-DBR.

• When S-DBR would not fit?

• Just a little about how to cause the change.


Learning from the development Learning from the development of conceptsof concepts

TOCICO 2009 Webcast

pp

• As TOC has significantly progressed since its initial development in 1985 – there must be something we can learn from the way ideas were formed.

• That initial verbalization was largely based on the OPT thoughtware that was developed in the early 80s.

R li i ft i it lf d t h l ’− Realizing software in itself does not change people’s mind.

• Since then TOC has expanded its scope• Since then TOC has expanded its scope − TP, CCPM, Marketing and Sales and lately also Strategy

and the concept of the ever-flourishing organization.


p g g

Learning from the development Learning from the development of conceptsof concepts

TOCICO 2009 Webcast

pp

• Few years ago some important changes to TOC in Manufacturing were introduced.

− It is beneficial to understand the shift of ideas and especially the change in the focusespecially the change in the focus.

− The first shift, from OPT to DBR, happened before the TOC and the five steps were verbalized.

− Then buffer management was introduced.

− Then the DBR technique has shifted to S-DBR.

− Then came the development of make-to-availability.

− All those shifts can be characterized as becoming


simpler, and also by moving from detailed planning to setting the priorities right.

OPT was all about planning! OPT was all about planning! Execution meant simply following the scheduleExecution meant simply following the schedule

TOCICO 2009 Webcast

p y gp y g

• Planning means making decisions ahead of time.− Which makes some decisions more vulnerable to

Murphy.

There m st be a reason for a specific decision to be− There must be a reason for a specific decision to be taken at the planning session. − Most detailed plans suffer from frequent changes in the p q g

execution phase.

− So, what decisions should be included in the planning?

− and what should be left for the execution?


From OPT to DBRFrom OPT to DBRTOCICO 2009 Webcast

• DBR was developed as a departure from OPT, whichDBR was developed as a departure from OPT, which was an early and very sophisticated advanced planning and scheduling software.

− Eli Goldratt realized the need to be simple with very clear rationale.

The emerged understanding backed the idea that there− The emerged understanding backed the idea that there was no need for software.

• The detailed planning (finite capacity scheduling) wasThe detailed planning (finite capacity scheduling) was limited to just one resource.

− None of the other resources were scheduled.


From OPT to DBRFrom OPT to DBRTOCICO 2009 Webcast

• Buffers were added to protect the sequence of the p qcapacity-constraint-resource (CCR) and the timely shipment to the client.

− Three buffers were used:1. The CCR buffer protecting the detailed schedule of the

CCRCCR.

2. The shipping buffer protecting the delivery dates. The CCR schedule had to ensure there is enough shipping buffer.

3. The assembly buffer was required to ensure that materials that do not go through the CCR would be


materials that do not go through the CCR would be released on time.

The main insight of TOCThe main insight of TOCTOCICO 2009 Webcast

gg

Fluctuating market demand

A complicated shop floor

Raw materials List of orders

Murphy

The critical insight of TOC is that only ONE


The critical insight of TOC is that only ONE bottleneck or capacity-constraint-resource (CCR) really governs the performance of the whole system

The buffers in traditional DBRThe buffers in traditional DBRTOCICO 2009 Webcast

A complicated shop floor

Raw materials List of orders

The CCR buffer The shipping bufferpp g

The assembly buffer


The DrumThe Drum--BufferBuffer--Rope (DBR) Rope (DBR) planningplanning

TOCICO 2009 Webcast

p gp g

• The Drum is the schedule of the CCR based on theThe Drum is the schedule of the CCR based on the orders at hand.

• Buffer is a protection mechanism to keep the p pdrum from disruptions.

− Buffers became a critical part of the planning.

• Rope is the raw materials schedule – forcing the choke of the release.

− Preventing too much WIP.

• All in all DBR represented a huge simplification f th l i


of the planning.

The appearance of Buffer The appearance of Buffer ManagementManagement

TOCICO 2009 Webcast

gg

• In the early days of DBR there were no guiding rules for the execution phase.

• Then Eli Goldratt developed the idea of Buffer Management to dictate the priorities in executionManagement to dictate the priorities in execution.

− Monitoring the actual consumption of the time buffer.− Dividing it to the three colors: Green, Yellow and Red.g ,

− Management should intervene only to flow red orders out.

• It took time for us to understand the contribution of B ff M tBuffer Management.

− Especially its impact on the distinction between planning and execution.


p g

Two conceptual challenges at Two conceptual challenges at traditional DBRtraditional DBR

TOCICO 2009 Webcast

1. Should the constraint lie in the market demand or inside the organization?inside the organization?− It was assumed we cannot have both at the same time.− At the time the thinking was that the strategic constraint g g

is under better control when it is inside the organization. 2. Suppose we do have an active CCR – is there a need

to schedule it in detail?to schedule it in detail? − What is the damage if the sequence is not kept?


The relationships between the internal The relationships between the internal capacity and the market demandcapacity and the market demand

TOCICO 2009 Webcast

p yp y

• The price we pay for having an internal capacity constraint:

− Delivery times are relatively long.

− Because of the wait time at the CCR.

− Due-date performance is threatened.

− Any interruption on the CCR itself puts the rest of the orders in the queue at a certain risk.

− Fact of life: clients are not interested that their supplier has a bottleneck - when they feel their deliveries are compromised they look elsewhere.


deliveries are compromised they look elsewhere.

Changing the focus to be on the demand Changing the focus to be on the demand ––the immediate ramificationsthe immediate ramifications

TOCICO 2009 Webcastthe immediate ramifications the immediate ramifications

• Focusing on the market demand is target to achieve the following:

− Expanding the current (potential) demand.− Charging higher price for higher value.

− For instance, by delivering faster and/or more reliably.− Making the shorter and reliable delivery valid at all times!− Making the shorter and reliable delivery valid at all times!

• Being able to focus on the demand requires protective capacity even on the “weakest link” itself!y

− Actually it means the CCR has to subordinate to the market demand!

M ti th it t t ll ti


− Meeting the commitments at all times.

Challenging the technique of Challenging the technique of traditional DBRtraditional DBR

TOCICO 2009 Webcasttraditional DBRtraditional DBR

• The damage of scheduling the CCR:g g− Inflexibility in dealing with urgent requests and other

changes in priorities.− Rescheduling causes changes to many orders.

− The scheduling process itself is batch oriented.O d it t b h d l d− Orders wait to be scheduled.

− Scheduling the CCR for make-to-availability is especially damaging because of the inflexibility it imposes.g g y p

• Maintaining three different buffers is difficult to implement and it creates conflicts in priority.


Challenging the technique of Challenging the technique of traditional DBRtraditional DBR

TOCICO 2009 Webcast

• What prevents us from deciding upon the exact i f th CCR i th l t i t ?sequencing of the CCR in the last minute?

− We need to know when to release the orders.− Providing enough time to be delivered on time− Providing enough time to be delivered on time.− Providing enough time to ensure the CCR is not starved.

− We like not to waste too much capacity of the CCR.We like not to waste too much capacity of the CCR.− Do we lose the precious capacity of the CCR if the CCR

deviates from the sequence?− Only when the CCR is subject to strong sequence-

dependent-setups there is a waste of capacity of the CCR.

• Note, the CCR-Buffer protects the sequence on the CCR


Note, the CCR Buffer protects the sequence on the CCR – not just preventing starvation.

The direction of the solutionThe direction of the solutionTOCICO 2009 Webcast

• The necessary ingredients of the solution for make-to-order:

− The planning is able to link between the need to meet all the commitments to the market and having enoughthe commitments to the market and having enough capacity of the CCR.

− There is one time buffer per order.p− The focus in the execution phase is definitely on the

market demand.

− There is one clear list of priorities based on the color code of green-yellow-red.


The Planned Load The Planned Load -- The tool for The tool for checking the capacitychecking the capacity

TOCICO 2009 Webcastchecking the capacity checking the capacity

• The following concept of DBR is still a cornerstone in the proposed S-DBR solution:

− Complex as the operational system may be, the output is dominated by one resource with the leastoutput is dominated by one resource with the least capacity.

• Based on that concept, the capacity monitoring,

D fi iti Pl d L d

replacing the detailed schedule, is based on the following concept:

Definition – Planned Load:

The Planned load is the accumulation of the derived load onthe CCR of all the firm orders that have to be delivered


the CCR of all the firm orders that have to be deliveredwithin a certain horizon of time.

A visual representation of the A visual representation of the planned loadplanned load

TOCICO 2009 Webcast

pp

rder

rd

er 5

5 --1

0 10

ours

ours

Daily Load 0

hour

s0

hour

s

Ord

er

Ord

er 3

3 --8

8 ho

urs

hour

s

rsrsO

rO

rhoho

21

21 h

ours

hour

s

urs

urs

(Hours)

Ord

er 1

O

rder

1 --

2020

r r 2 2 --1

4 14

ss er

er 4

4 --2

0 20 h

our

hour

Ord

er

Ord

er 6

6 --22

er

er 7

7 --1

8 18 h

ouhou

Ord

erO

rder

hour

sho

urs

Ord

eO

rde

Ord

eO

rde


Time (days)

The critical information obtained The critical information obtained from the plannedfrom the planned--loadload

TOCICO 2009 Webcast

pp

• The most important feature of the planned load is:A rough estimation of the time for a new order to wait

until being processed by the CCR

• This time estimation allows us to determine a “safe-date” for which we can commit to complete the new order.

We need to add certain time to the planned load to− We need to add certain time to the planned-load to provide the order with enough time to be processed by the CCR and the downstream operations.− Considering also possible delays and flexibility given to the

CCR in deciding when to process that order.

− The default in S-DBR is to add ½ of the production buffer


The default in S-DBR is to add ½ of the production buffer time to the planned load to determine the safe date.

Using the plannedUsing the planned--load to estimate safe load to estimate safe delivery dates for new incoming ordersdelivery dates for new incoming orders

TOCICO 2009 Webcast

All the required CCR load for the ordersAll the required CCR load for the orders that have a committed delivery date have been accumulated here.

DailyLoad

We cansafelydeliver here

50% of the production buffer

deliver here

The planned load

Time (days)


The subtle difference between the planned The subtle difference between the planned --load and the detailed schedule of the CCRload and the detailed schedule of the CCR

TOCICO 2009 Webcast

• The planned-load looks like a schedule because it l th i d CCR’ ti th ti fplaces the required CCR’s operations on the time frame.− But, it does not impose the sequence on the operators. − Many times the time allotted to an order on the planned-

load is not realistic.− For instance, when several different operations require the

CCR for the same order.• All we need the planned-load for is to estimate the

earliest due-date for a new order that is pretty certain.− And for that the rough estimation is enoughAnd for that the rough estimation is enough.− Having capacity requirements and response time

together creates an effective tool for monitoring capacity.The overall buffer time provides enough flexibility to


− The overall buffer time provides enough flexibility to maneuver the actual sequence on the CCR.

Sales and Production planning in Sales and Production planning in SS--DBRDBR

TOCICO 2009 Webcast

• The planned-load is used to support decisions on itti t li bl d d tcommitting to reliable due-dates.

− Not necessarily Sales should always offer the safe-date.At k ti th f d t t l d t th l− At peak-times the safe-dates get longer due to the longer queue of work already committed to be delivered.

− At off-peak times the safe-dates are relatively early. p y y• The visibility of the “safe-dates” should be easily

accessed by the sales people as the critical information i d f P d tirequired from Production.


The advantages of SThe advantages of S--DBR over DBRDBR over DBRTOCICO 2009 Webcast

• The managerial focus is put mainly on the market d ddemand.

• Sales are properly supported online by giving a direct access to “safe-dates” for any new order.access to safe dates for any new order.

• Buffer management in S-DBR is much more effective in handling the priorities than the three separate buffers in DBR.

• S-DBR can be implemented faster –implemented faster bring results sooner.


The advantages of SThe advantages of S--DBR over DBRDBR over DBRTOCICO 2009 Webcast

• S-DBR requires much simpler algorithm from the ti ft th DBRsupporting software than DBR.

− To illustrate the complexities DBR sometimes faces:Scheduling several CCR operations for one order− Scheduling several CCR operations for one order.

− Scheduling several non-identical units of the CCR.

− Scheduling resources like ovens which take several ordersScheduling resources like ovens, which take several orders and work on the whole batch rather than one piece at the time.

All the above do not complicate the S DBR algorithm and− All the above do not complicate the S-DBR algorithm and it remains effective.

• When the constraint moves to another resource, S-DBR


,needs only to change the ID for the planned-load.

When SWhen S--DBR cannot applyDBR cannot applyTOCICO 2009 Webcast

pp ypp y

• The idea of refraining from detailed schedule of the CCR and yet being able to control the CCR’s capacity is basedand yet being able to control the CCR’s capacity is based on the assumption:

The sequence of production does not significantlyThe sequence of production does not significantly impact the capacity.

• The one situation where the above assumption is not valid is called: sequence-dependent setups.

− The setup time heavily depends on both the previous setup and the next onesetup and the next one.

− This situation calls for having to follow a “preferred sequence”, which means very limited ability to expedite.


y y

− Handling this situation is beyond the scope of the webinar.

Implementing SImplementing S--DBRDBRTOCICO 2009 Webcast

p gp g

• Important insight regarding implementation of a paradigm change is:paradigm change is:

It is imperative to get results very soon!− This means we must plan the implementation so the p p

first steps will already generate significant results.− There is definitely no need to identify the weakest link /

CCR at the start of the implementationCCR at the start of the implementation.• The first and most important action has to be: choking

the release.− Limiting the WIP to contain only what has to be

delivered within the production time buffer.• The default for the initial production buffer is: 50% of


• The default for the initial production buffer is: 50% of the current production lead-time.

DBR from a different perspective: DBR from a different perspective: guidelines for ANY planningguidelines for ANY planning

TOCICO 2009 Webcast

g p gg p g

• Drum: Identify and carefully plan the critical areas where any deviation would clearly cause damage (to the objectives of the planning).

• Buffer: Protecting the drum instructions mainly from• Buffer: Protecting the drum instructions mainly from disruptions at the non-critical areas.

• Rope: Protects the buffers from over-protection. − Usually is means choking the release of work/missions to

ensure adequate protective capacity.• Note protective capacity is part of the overall buffering• Note, protective capacity is part of the overall buffering

system.• The above is true for S-DBR, CCPM and for any type of


planning.

The road to simple and effectivesolutions

TOCICO 2009 Webcast

• The concepts for managing manufacturing e co cepts o a ag g a u actu gorganizations were not invented in very short time.

• Coming out of complexity to recognize the simple rules behind it is not easy.

• Once we recognize the distinction between planning and execution we move to a simpler yetplanning and execution we move to a simpler, yet effective, scheme.

• And the basics of TOC still hold!And the basics of TOC still hold!


Questions?Questions?Questions?Questions?

TOCICO 2009 Webcast

Eli Schragenheim’s new book, together with Bill Dettmer andtogether with Bill Dettmer and Wayne Patterson:Supply Chain Management at W S dWarp SpeedWill be out in May 2009Covering the new knowledge in

Eli S h h i

Make-to-order, Make-to-Availability and Distribution.


Eli [email protected]

from dbr to simplified-dbr-8 (2) - tocico.nettocico.net/from dbr to simplified-dbr-8 (2).pdf · the...

Documents