Week 3 -Black Belt Training

Below is a review of week three of Lean Six Sigma Black Belt training, this blog is intended to assist students in thier study of this subject.

Why do we use regression/simple linear or multiple? To evaluate the relationship between leading indicators (Xs. KPIVs) and the output the customer receives (Ys, KPOVs). PG 58-59

In regression/simple linear or multiple - What kind of data is needed for X and Y? Both X and Y need to be continuous variables. PG 67 and 68

Why do we use multiple regression instead of simple linear regression? We believe that more than one X impacts the Y significantly. We want to evaluate the impact of the combination of Xs on the Y. PG 58

Once we have run a regression model, what is the first thing we need to check? Residuals PG 69

Multiple Regressions – What are we looking for in each graph (1) Normal distribution of residuals (2) Randomness no pattern in fitted value (3) Centered about 0 randomly above 0 line
PG 70

Multiple Regression – Bob the BB is working on a process. He and his team believe that there are 5 inputs that predict the output. Will the regression model that uses all 5 Xs have the most explanatory value for Bob and his team? Maybe Yes Maybe No – 2 or more of the Xs might show evidence of multicollinearity - Bob and the team must investigate.

What technique will Bob and the team use to determine the best model?
Best Subsets
PG 71

In Best Subsets what are the terms used to evaluate the models identified by Minitab?
R2 - high
Adjusted R2 - low
MSError (small MSError is desired)
Mallows C-p statistic (small C-p desired, close to the number of terms in the model)
PG 72, 74, 75

Should Bob and the team use a one factor model – Why?
No the R2 (adj) for both one factor models is less than one of the 2, 3, 4, factor models.

Should use the 2 factor model with X1, and X4 – It has the highest R2 (adj) lowest mallows CP and lowest S
PG 75

For ANOVA what type of data is needed for X?
(X) discreet (Y) continuous

In ANOVA (and hypothesis tests for that matter) what is the null hypothesis (Ho) – There is no difference – all of the means are equal. PG 102

When using an ANOVA what are we trying to determine? We want to know if a change in a discreet X has a statistically significant impact on the output of Y. PG 141

Why don’t we just use a box plot? A box plot is an excellent graphical tool but it is of limited value when the data is close together. PG 99

What is a degree of freedom (DF)? A statistical credit (currency) equal to one less than the number of elements studied. It is equal to N-10 where N is the number of elements. PG 122

In the ANOVA how many data points were evaluated?
Let’s say we are testing a factor that has five levels and we collect seven data points at each factor level…
How many observations would we have? 5 levels x 7 observations per level =35 total observations
How many total degrees of freedom would we have? 35 -1 = 34
How many degrees of freedom to estimate the factor effect? 5 levels -1 = 4
How many degrees of freedom do we have to estimate error? 34 total -4 factor = 30 degrees of freedom

Is the between greater than within? Variability mean (squared) yes the facility is greater than the error in MS -- signal to noise. PG 103

Yes the mean square for the facility (between) is larger than the mean square within.

How much bigger is it? 193.86 times (f, ratio) = MS Facility/MES = 171.726/0.886 the signal is 193.86 times louder than the noise (Matilda)

Is that big enough? Yes because the p value is 0.00 which means a very, very small chance of error from rejecting the null hypothesis. (Needs to be less than .05 for P Value)

What are the 3 steps along the path to eliminating the cost in any organization:

Develop fact base (complexity diagnosis)
Implement improvement (reduce # of requirements, improve speed, reduce offering)
Attack creators of complexity (remove root cause)
PG 151

What tools can be used to reduce complexity in any organization?
Lean Tools
PG 157

3 factors contribute to queues?
(1)Average service time
(2)Utilization
(3)Variability of Service and Arrivals
PG 164

What happens as resource utilization increases from 70% to 100%?
The queue will increase exponentially

Why do we need solution development in the improve phase?
While we know show our Xs impact our Y we haven’t yet sold others on the solution to our problem statement. Capture out of the box ideas and increase chances of solution acceptance.
PG 197-198

Idea generation methods ~
Brainstorming
Random Word
Personality Challenge
6 thinking hats
20 questions
PG 201-205

How might we prioritize and/or force rank the ideas generated?
Nominate Group Technique
Benefit/Effort Matrix
Criteria Based Matrix
PG 207, 208, 209

What is a 2 Bin System?
A simplified version of a standard replenishment pull system and maintains the benefits:
Establish Buffers
Decouples Supply Process
Triggers Based on Consumption
PG 228

What does a Service Level 1 mean when calculating safety stock?
On average we would eliminate 84% of the time
PG 221
What does safety stock guard against?
Demand Variability
Lead Time variability
Quality Variability

Part JMB-2006 exhibits the following characteristics:
Order Interval (01) 10 days
Lead Time (LT) 5 days
Demand (DMD) 100 parts per day
Safety Stock (SS) 5 days

What is the Max Loop?
2,000 parts

What is the trigger point?
1,000 parts

What is the average on hand?
1,000 parts
PG 225

What is the 4 step method of Streamlining?
Value Add Analysis is always first
If the process is not meeting customer demand – process constraint identification
If the process is meeting customer demand but the tasks/work loads are not balanced-process balancing
If the tasks are balanced but there is too much movement of information material – process flow improvement
If the process has limited movement but has too much non-value add time – value add improvement

Process Constraint Identification
A process can have one or more time traps – True
Every process has a constraint – False
Every constrain is a time trap – True

Maggie owns a hamburger joint. Her average demand for double cheeseburgers is 300 per day. The shop is open 9-11. Takt time –
300 burgers per day/10 hours per day = 30 burgers
# of units required/net available time
Takt time – every 1 hour = 30 burgers
1/30 30/60 invested takt rate
The takt time is 2 minutes per burger

What is process balancing?
A procedure where the process steps are equalized in terms of time required to accomplish them.
PG 275

To improve the flow of the process –
Collocation of personnel
Mobile equipment
Digitalization
PG 301

4 Techniques to improve process flow –
Remove Non Value Added (NVA)
Improve layout
Focus workspace
Rebalance tasks
PG 302-303

What process map shows the need for flow improvement?
Spaghetti
PG 298

4 step value added improvement method:
Document Process Steps and Separate Activities into Essential (CVA) & Auxiliary (BVA/NVA)
Eliminate or Streamline Auxiliary Activities
Streamline Essential Activities
Improve Equipment / Service Design to speed up remaining activities
PG 317