Critical chain buffer Management

Critical Chain

Introduced by Goldratt

• Improved project success

• Reduced project duration

• Increased project team satisfaction

• Simplified project measurement

• Simplified project management

• Increased project throughput with same resource - resource efficiency increased

Critical chain vs theory of constraints

Critical chain is nothing more that theory of constraints applied on project management

Operations Management (Theory of Constraints):

Target: throughput - produce as much as you can per time

Constraint: The bottleneck machine determines the maximal production quantity

Remedy Solution - inventory buffers (add some inventory so bottleneck machine would not delay the process)

Critical Chain:

Target: Time

Constraint: Critical chain - determines the total project duration

Remedy: Time buffers

Critical Chain/Buffer Management CC/BM planning methodology

1. Come up with aggressive estimates

2. Construct an ALAP schedule - as late as possible

3. Identify the critical chain

4. Determine appropriate buffer positions (in the end)

5. Detemine appropriate buffer sizes

6. Insert the buffers into the schedule

Come up with aggressive estimates

The vicious cycle:

we’re scared things will go wrong → we add extra safety time to estimates

More trust - more relaxed

Less attentive

Consumption of safety time

Exceeding activity durations

Return for more safety time

If something can go wrong, it will anyway

So, Instead of letting every person hide safety time in their own task, pool it into shared buffers — and actually manage those buffers actively.

People tend to add protection time to their estimates to feel safe, which ends up being equal to 90% confidence level. CC/BM says to take the average (median) estimate instead — the 50% confidence level.

Step 2. Construct as late as possible schedule

Work backwards from project deadline

• Minimize work-in-progress (levelling resources)

• Delay important cash outflows

• Less rework (more info you get → less rework)

• BUT all tasks become critical (!no slack left)

Prons and cons with Goldratt Critical Chain Method

Advantages:

• It is a new technique - some of the software offers the solutions with those methods

• Buffer insertion

Disadvantages

1. Critical chain depends on the method used or objective selected, so identifying critical chain is a difficult part, it is unique, insertion of buffers changes the CC, quality depends on the CC

2. Buffer insertion - 50% rule is a bad random rule, better based on risk and resource scarceness

3. Activity estimates - tricky on how to convince people to be honest with you. Bottom-up instead of top down

4. Roadrunner mentality - not clear dates for subcontracters, so more subcontracters →weaker the technology is

5. Buffer Management — You need to constantly track how much "safety time" you've used up. People usually use the major part of the buffer at the beginning of the project.

6. Multi-project — If people work on multiple projects at once, everything slows down because they're always waiting for each other.

Step 4. Determine appopriate buffer positions

Collect all buffers you took away from each activity and put 50% of the critical chain amount in the end.

Risk (variability) - sometimes you can finish earlier, other time later - those things cancel out, less buffer needed

As a result - can promise much quicker project deadline for a client then with Regular technique

Roadrunner mentality - no longer milestones, as you finished earlier → other person starts earlier, finish later → starts later. Not based on milestones that extend the deadline

Size of the buffer two methods

• 50% of critical chain - quite random but simplified

• Sum of squares method - proportional to the riskiness

• Adaptive method with network density

• Adaptive method with resource tightness

Buffer - protects only the risk of delays of activities on the critical chain!!!

Feeding buffer

Add to any feeding chain that enters the critical chain

The feeding chains - the sub network of activities not lying on the critical chain, non-critical activities lie there

Project buffer - protects the project duration against delay in the activities of the critical chain, onw project buffer

Feeding buffer- protect the critical chain from the delays in activities in the feeding chain, multiple feeding buffers - in between the end of feeding chain and a critical chain

Resource buffer

Placed when a resource is needed on the CC, and the previous CC activity will be performed by another resource type

placed before a critical chain activity that needs a scarce or important resource.

Action threshold

At the green zone - everything goes well

Orange zone - watch out

Red zone - almost all time buffer is consumed, take actions

Green zone increases as the project reaches its end, so in the end less risk (if you consumed almos all your buffer in the end - no worries as your project has almost finished)

Root squared error (RSM)

Bigger risk → bigges size of the buffer

The buffer size is equal to the square root of the sum of squares of differences between the normal duration and aggressive duration

√((normal1 - aggressive1)² + …(normaln - aggressiven)²)

Adaptive method with network density (AND)

Buffer depends on risk and network structure

difference = normal - aggressive

√((difference1*0.5)² + …(differencen)²) * (1 + network density)

Network density = number of arrows btw activities / number of activities

Problem with adding buffers

If you add buffer → new resource conflicts

BUT after solving → hidden buffers appear

ART (Adaptive method with resource tightness)

Considers for 50% of the riskiness + resource criticality

ART = (1+Resource tightness)*RSM(50%)

Resource tightness = total work content used/total work content available

Total work content used - the activity on a chain with the largest work content

Total work content available = sum of the work content of activities on a chain