Learning Curves

Learning curves: please note that this lecture relates to Chapter 12 of the Course Notes (and not Chapter 11 as stated in the lecture)

1. says

at first i thought we need to calulate the total time for the 30th batch and then 31th batch

2. says

Thank you for the useful lecture. I was just wondering if u can help me with working it out on the calculator as I can’t seem to find the symbol Xy and +/y to work out the answer for Example 7?

Many Thanks

• says

I can’t really help, because there are different types of calculator with different ‘logics’ which affects the order in which you press the keys.
Sorry, but you are going to have to search in the manual that came with the calculator

• says

I thought you might say that I have found how to do the Xy symbol but not the +/- but I’ll have a look at the manual again. Thanks again for the lecture!

3. says

Hello Sir, thanks for the great work. I need a clarification. Why is it that when learning rate increases, average time also increases. I thought as learning rate increases which means people learn faster, then average time taken to produce the product should reduce. For example when first unit takes 200hrs and learning rate is 60% then average time is 120hrs for second unit, if the rate increases to 70%, the average time rather increases to 140hrs which confuses if compared to real life.

• says

I realise that in other contexts we tend to think of it as the higher a percentage it then the better, but in the case of learning rates this is not the case.

The lower the learning rate the faster they are learning. We would want the learning rate to be as low as possible. A learning rate of 70% is better than a learning rate of only 90% – the 70% people are learning faster.

• says

Thanks for the quick response but still not very clear. What does the percentage really mean then? I take it as it means the rate at which the workers are learning and if so then the faster the better so the higher the rate the better. That’s why am thinking that way. Probably it means something different. I will be clear if u can help me with that thing.

• says

Don’t look for a meaning! All it means is that the average time falls to that percentage of the previous average time.

It is a measure for us to find how fast they are learning, but not in the way that you are trying to interpret it. Too many think that way without understanding how it is working (which is why the examiner once asked as a small part of a question – which is better, a learning rate of 80% or a leaning rate of 90%. The answer is 80%).

(Higher percentages are not always better. If I told you that my company wastes 10% of my raw material and another company wastes 5% of its material, it certainly does not mean that my company is doing better )

• says

I got it. Thanks very much. It means For example, an 80 percent
learning curve means the per unit average
cumulative cost (in hours or dollars) falls to 80
percent of the previous per unit average
cumulative cost as the cumulative output
doubles. Therefore the lower it falls the better.

4. says

Hi John,

Just a note that “r” is defined incorrectly in the notes. It should be “…expressed as a decimal” as opposed to a “…%”.

Thank you for a great lecture as always!

5. says

Sir, how can i determine the learning rate in a learning curve question if it was not given in the question?

• says

The learning rate will be given in parts of questions where you need it.

As a separate part of a question you could be told how long each of the first and second units/batches take, and then you can work out the learning rate ‘backwards’ using the doubling rule.

• says

It is meaningless because it is impossible.

(Just think, how on earth could the average time per unit be zero – unless the time for the first unit (and all subsequent units) was zero!!)

6. says

Thank you sir, this lectures allowed to understand the logic behind the learning curve formula.

7. says

Hi all,

Could anyone let me know how to use Casio scintific calculator for learning curve formula..say

Y=aX^b = 10*5^-0.152

Thanks
RB

8. says

Thank you so much. I had problems understanding this before, but now is so clear.

Regards

9. says

great lecture sir. Can you pls advise if we should emphasise only on theory part for Regression Analysis & Time Series or should we also concentrate on calculatios equally.

thanks

• says

As is made very clear in the introduction paragraph to the Course Notes, you will not be asked for calculations on regression or on time series. (The calculations are examined in Paper F2)

You are only expected to know the principles. I have only left the calculations in the notes (and lectures) to hopefully make the idea more clear.

10. says

great lecture sir. Can you ls advise if we should emhasi only o theory art for Regression Analysis & Time Series or should we also cocetryae o calculatios wqually.

thaks

11. says

Dear Sir,
I read in the syllabus for Dec 2013, that the only topic left in the syllabus is learning curve and high-low method in quantitative budgeting…is it?

Many Thanks

• says

Time series and regression analysis are assumed knowledge from Paper F2. The examiner has said that there will not be detailed calculations on either, but you are expected to be aware of both.
(If you look at the course notes, it actually says this in bold letters. We update our notes every six months and everything in our notes is relevant for the December 2013 exams)