Thanks a lot, i'm going to use AHP in my graduation project in risk management.. your video is simple and easy to follow. i'll let you know how things will end with me.
Excellent clip , I highly appreciate if you could elaborate more on how to relate AHP to financial parameters such as Net present value (NPV) or internal rate of return (IRR)?
Is it possible to upload a complete template to illustrate your video? I have tried to enter the numbers you used and am getting slightly different results. For instance, while you demonstrate values of 17%, 43%, and 40% weights for the color, memory, and delivery criterion, respectively. ... I obtain 16.9%, 44.3%, and 38.7%. A completed model may help me better understand you input method and results. Thanks much for you contribution to this interesting topic (AHP).
I do not understand the basic principals of this tecnique. First we assume linear relationships between all criterias at the beginning. And then we give some value for all raitos of pairs. However from mathematical point of view, it is impossible to give ratio to all of them. For example, at the 4:32, we assumed that memory is three times and delivery is two times more important than color. But since we assume a simple linear relationship between all pairs, we can calculate the memory/delivery ratio as 3/2. How come we can assume that memory and delivery has the same importance in this case?
AHP allows inconsistency in judgments. It’s measured with the consistency ratio CR. The method works with “near consistent” matrices only, I.e. CR < 10%. Read the original paper from Saaty.
Assuming that the ratio of importances is a constant value is the same thing as linear relationship. I can understand that sometimes it is impossible to make accurate measurements especially in social sciences and important to allow some uncertainties but 33% (1 instead of 1.5) is too much. More importantly, does the method say anithing about errors or uncertainties?
Most users of AHP do not consider uncertainties, just inconsistencies. My software implementation gives uncertainty estimations due to rounding, e.g. 1 and 1.5
Ok. Thank you for quick replies. But my problem is simply related to pure mathematics behind it. What AHP does, as far as i understand, is to transform a desicion making problem to a linear eigen value problem at first. But when defining a 'ratio of importance matrix' some ratios can be easily calculated from the previous ones (again since we are interested in a linear problem). Only thing i'm having hard time is that point.
Thank you for the video but I am yet to understand how to input the answers gotten from the respondents. In a situation where I have over 100 respondents, what is the way to go?
Have a look at bpmsg.com/ahp . There you find an AHP online calculator and the free AHP online SW for more than 1 participant. You will see, how the decision matrix is filled with the judgment values. For consolidation of more than one participant I am using the geometric mean - Aggregation of individual judgments (AIJ). All mathematical relations are shown in my published paper: doi.org/10.13033/ijahp.v10i3.590
You have to multiply the "local" priorities with the priorities one category (level) higher in the hierarchy, in order to get the "global" priorities. Example: pink = 13% * color = 17% = 2.2%
Very nice video. I have only one question. How this method works when we have sub-categories? For example the appearance of a cell phone may include: colour, size and other factors.
Yes, you are right. The correct numbers are 16.9%, 44.3% and 38.7% (color, memory, delivery), 13.6%, 9.8%, 4.3%, 26.5% and 45.8% (colors), 4%, 9.5%, 41.7% and 44.8% for memory and 48.1%, 40.5%, 11.4% for delivery. The first version of my excel template (April 2010) was not very accurate with the power method, sorry. Inbetween the template is very precise and correct. You can get it from my website bpmsg dot com. A new version with 10 criteria will be released soon.
Hello, could you please tell us the matrix values that you entered for the memory and delivery sub-criteria? I have written a program that uses the AHP and I would like to test it using your values. Thank you!
BPMSG But that only has the final percentage values. Could you please give us the square matrix values (like at time 6:00 in the video, but for memory and delivery)? Thank you so much!
ANP is more general than AHP and also complex. I still need some study to fully understand, and find a way to present the essentials. Will take time, but plan to upload once it's done.
Respected sir, I want to use AHP on my research work but I am little confused about its application. Actually I am working on inspection skill of human labour and factors affecting on it. And I want to use AHP to classify the available manpower into three skill levels (Low, Medium, and High) based on the factors that can affect the skill of quality inspectors. Plz guide me can i use AHP for this purpose? if yes then plz give your valuable suggestions and useful material that can be helpful for me to solve my problem.
Thank you so much for this wonderful tutorial. Please, I want to use AHP for risk assessment, any idea of the materials I can use for quick reference? Thanks.
Thank you for your clear explanation. I have a doubt: I noticed that in many cases AHP is used for defining sub-criteria priority even though they are already "numerical". EG, In the example you did, the "memory" criteria is already numerical: an option could be to normalize those values in some mathematical way (linear, log, etc) and then use these vector instead of apply AHP. Of course, in this case, I mean that the user of the example think that there is no real difference between 32 or 64 MB, but in this case there were only 4 alternatives. In many application there are a lot of alternatives (for example, if applied in a raster image), so pairwaise comparison would be impossible to do. In many cases I saw AHP application on numerical criteria which are actually far more complex than use the original data (once normalized, of course). Don't you think that in many cases AHP is misunderstood? It seems to me that the strenght of this method is its simplicity, but in many cases people like to complicate their lives! Thanks again for your video and for your site too.
You are right, it is also possible to use numerical values on a ratio scale instead of doing the pairwise comparisons, and it is done in AHP projects. My example just illustrates that the perceived value of memory could be different from the numerical ratios for a decision maker. Anyway, you run into problems with AHP if the number of alternatives exceeds 7 +/- 2. If you have a lot of alternatives, you can use AHP for the criteria weights only, and combine with another method for alternative evaluation, for example it could be a simple rating table.
Perfectly clear, thanks for your point. I saw many AHP projects where there are some confusion about the real aim of the method and how to use it and that leaded to misleading results. I think that AHP is a flexible option to make your choise easier, but in some case is used in a very rigid way and make things really complicated. So thanks again for the useful example of the video and for your comment. Best regards.
+Kanyarat Phudphad You could map 1 2 3 4 5 to the AHP scale 1 3 5 7 9. Important is that your questions are pairwise comparisons, i.e. "what is more important, A or B, on a scale 1 to 9 (in your case 1 to 5)
Goodmorning! BPMSG Can you help me. Do you have a sample program for AHP (Analytical Hierarchy Process) that can help for my study. Thank you in Adnvance.
The excel template can be unprotected w/o password. You can see all formula. My AHP online software is not open source, but all mathematical relationships are shown in the manual (you can download from my website) Regards, Klaus
I don't know a specific address to find something abt AHP and marketing. Just google. There is one free online journal on AHP. IJAHP.org. Maybe you find something there.
I'm so grateful watching this video, Mr. Klaus. Thank you so much, some literatures and journals have made me confused.
this is the best video to understand the whole AHP process
Thanks a lot, i'm going to use AHP in my graduation project in risk management..
your video is simple and easy to follow.
i'll let you know how things will end with me.
This is by far the best explanation of AHP, thank you so very much!
I didn't have to look for another source after watching this video, be blessed Sir
Excellent video and step by step cleared out everything. Thank you
the most useful tools for decision making, many thanks
understood it very well - simple graphics, effective teaching. please make more videos !
Thank's a lot Mr. Klaus. It is so helpful to understand AHP.
Thank for sharing Mr. Klaus, have a good life!
Danke sehr! Sie retten mir meine Klausur!
A very concise and wonderful explanation
Amazing! Wonderful step by step tutorial. Thank you so much!
Fabulous video Mr. Klaus! It helped me and my team a lot!
Thank you very much! Great explanation and video. Also, thanks for your online tool. It's awesome and concise!!
Very good explanation, this seems like a very useful practical tool for making rational decisions.
Hello! I didnt understand how have you calculated the weight 17,43 and 40
Excellent clip , I highly appreciate if you could elaborate more on how to relate AHP to financial parameters such as Net present value (NPV) or internal rate of return (IRR)?
Wonderful explanation of AHP - thank you Klaus!
i know Im kind of randomly asking but do anyone know a good website to watch newly released movies online?
Great explanation, concise and clear, Thanks a lot
Wow you save my reading time. Thanks a million.
Is it possible to upload a complete template to illustrate your video? I have tried to enter the numbers you used and am getting slightly different results. For instance, while you demonstrate values of 17%, 43%, and 40% weights for the color, memory, and delivery criterion, respectively. ... I obtain 16.9%, 44.3%, and 38.7%. A completed model may help me better understand you input method and results. Thanks much for you contribution to this interesting topic (AHP).
Can we use Likert Scale for Analytical Hierarchy Process to obtain weights?
I do not understand the basic principals of this tecnique. First we assume linear relationships between all criterias at the beginning. And then we give some value for all raitos of pairs. However from mathematical point of view, it is impossible to give ratio to all of them. For example, at the 4:32, we assumed that memory is three times and delivery is two times more important than color. But since we assume a simple linear relationship between all pairs, we can calculate the memory/delivery ratio as 3/2. How come we can assume that memory and delivery has the same importance in this case?
Where do we assume linear relationships between all criteria?
AHP allows inconsistency in judgments. It’s measured with the consistency ratio CR. The method works with “near consistent” matrices only, I.e. CR < 10%. Read the original paper from Saaty.
Assuming that the ratio of importances is a constant value is the same thing as linear relationship. I can understand that sometimes it is impossible to make accurate measurements especially in social sciences and important to allow some uncertainties but 33% (1 instead of 1.5) is too much. More importantly, does the method say anithing about errors or uncertainties?
Most users of AHP do not consider uncertainties, just inconsistencies. My software implementation gives uncertainty estimations due to rounding, e.g. 1 and 1.5
Ok. Thank you for quick replies. But my problem is simply related to pure mathematics behind it. What AHP does, as far as i understand, is to transform a desicion making problem to a linear eigen value problem at first. But when defining a 'ratio of importance matrix' some ratios can be easily calculated from the previous ones (again since we are interested in a linear problem). Only thing i'm having hard time is that point.
Please sir, can I have the link of where to download the Excel sheet to compute the normalized principal Eigenvector of the matrix. Thank you so much.
bpmsg.com under tools. ahp.bpmsg.com is a free AHP online software for more complex projects.
How did you calculate the relative cost in %?
Total costs of all alternatives are 120 +120 +150 +150 =540. Rel. cost for model of 120 = 120/540 = 22%
Thank you so much
Thank you for the video but I am yet to understand how to input the answers gotten from the respondents. In a situation where I have over 100 respondents, what is the way to go?
Have a look at bpmsg.com/ahp . There you find an AHP online calculator and the free AHP online SW for more than 1 participant. You will see, how the decision matrix is filled with the judgment values. For consolidation of more than one participant I am using the geometric mean - Aggregation of individual judgments (AIJ). All mathematical relations are shown in my published paper: doi.org/10.13033/ijahp.v10i3.590
Thank you very much! it helps me!
Hello from Timor-Leste
Very well explained. Thanks Klaus!
thanks for such wonderful explanation
You have to multiply the "local" priorities with the priorities one category (level) higher in the hierarchy, in order to get the "global" priorities. Example: pink = 13% * color = 17% = 2.2%
Where is the mentioned Excel sheet? (4:56)
Can you post it, please?
Check bpmsg.com under tools.
Thank you so much .
Very nice video. I have only one question. How this method works when we have sub-categories? For example the appearance of a cell phone may include: colour, size and other factors.
Use AHP-OS free online software to define hierarchies: https:bpmsg.com/ahp/
Excellent video. Thank you, Sir.
Thank you very much sir, you are an excellent teacher..
Is it possible to write down the literature which you used for the video? Thank you very much for the explanation!!!
Yes, you are right. The correct numbers are 16.9%, 44.3% and 38.7% (color, memory, delivery), 13.6%, 9.8%, 4.3%, 26.5% and 45.8% (colors), 4%, 9.5%, 41.7% and 44.8% for memory and 48.1%, 40.5%, 11.4% for delivery. The first version of my excel template (April 2010) was not very accurate with the power method, sorry. Inbetween the template is very precise and correct. You can get it from my website bpmsg dot com. A new version with 10 criteria will be released soon.
hello i want to ask how you get the weights of sub criteria from main criteria as you have told at 6.16
17% = 17/100 =.17
13% = 13/100 = .13
.17 X .13 = .02 which is 2.2%
Sir, How did you arrive to the 17% Please assist. I tried to input in the tools as directed but with no luck.
Very good & explanatory video....Thank You... :)
Hello, could you please tell us the matrix values that you entered for the memory and delivery sub-criteria? I have written a program that uses the AHP and I would like to test it using your values. Thank you!
You could find the details on my website.
bpmsg.com/academic/ahp-group.php?sc=expl02
BPMSG But that only has the final percentage values. Could you please give us the square matrix values (like at time 6:00 in the video, but for memory and delivery)? Thank you so much!
Annjalie123 I posted all input data here: bpmsg.com/ahp-youtube-video-example-matrix-values-and-priority-vectors/
BPMSG Thank you so much!!
Sir what sub criteria u taken for delivery... If delivery time takes then more %will came... Then what should i do.... Which is best ???
Very well explained.
Very nice explanation. Thank you sir :)
Amazing explanation, thank you.
tq brother, how you get the percentage for pink= 2.2 %, blue = 2 % green = 0.9 % and so forth..Tq very much
ANP is more general than AHP and also complex. I still need some study to fully understand, and find a way to present the essentials. Will take time, but plan to upload once it's done.
Super duper explanation.
Great video!
Hi, thank you for an excellent video! If i want to learn more of how to use AHP in marketing research, where is the best place to search? :)
Hello, is it possible to run SEM for factors found by AHP?
I don't know enough about SEM (you mean simultaneous equations models or structural equation modeling?), sorry.
Sorry, don't know enough about it to answer your question. (SEM = structural equation modeling or simultaneous equations models?)
BPMSG Yes, I meant Structural Equation Modeling.
naser19851985 yes it can bro..some journals are available to combine SEM and AHP.
Very understandable intro thanks!
Respected sir, I want to use AHP on my research work but I am little confused about its application. Actually I am working on inspection skill of human labour and factors affecting on it. And I want to use AHP to classify the available manpower into three skill levels (Low, Medium, and High) based on the factors that can affect the skill of quality inspectors. Plz guide me can i use AHP for this purpose? if yes then plz give your valuable suggestions and useful material that can be helpful for me to solve my problem.
if there is equal importance on the criteria then how will i calculate? please inform me
sweety shahinur : then a “1” will be put into the decision matrix.
hi, where can I find the excel sheet? but thank you so much for the video. It is really helpful
bpmsg.com/new-ahp-excel-template-with-multiple-inputs/
bpmsg.com/ahp is free online software
Thank you so much for this video. would you like to share an example only with criteria?
You can check my website with the (free) AHP online software. There you find a couple of examples. bpmsg.com/academic/ahp.php
Thank you so much for this wonderful tutorial. Please, I want to use AHP
for risk assessment, any idea of the materials I can use for quick
reference? Thanks.
Have a look on my website. There is some material about AHP, as well as a free online AHP software to play around with: bpmsg dot com
thanks you so much, Can i get the calculation for up to 8 ?
Please check my website bpmsg.com There is an AHP excel template for download and a free AHP online software
Thank you for your clear explanation. I have a doubt: I noticed that in many cases AHP is used for defining sub-criteria priority even though they are already "numerical". EG, In the example you did, the "memory" criteria is already numerical: an option could be to normalize those values in some mathematical way (linear, log, etc) and then use these vector instead of apply AHP. Of course, in this case, I mean that the user of the example think that there is no real difference between 32 or 64 MB, but in this case there were only 4 alternatives. In many application there are a lot of alternatives (for example, if applied in a raster image), so pairwaise comparison would be impossible to do.
In many cases I saw AHP application on numerical criteria which are actually far more complex than use the original data (once normalized, of course). Don't you think that in many cases AHP is misunderstood? It seems to me that the strenght of this method is its simplicity, but in many cases people like to complicate their lives!
Thanks again for your video and for your site too.
You are right, it is also possible to use numerical values on a ratio scale instead of doing the pairwise comparisons, and it is done in AHP projects. My example just illustrates that the perceived value of memory could be different from the numerical ratios for a decision maker. Anyway, you run into problems with AHP if the number of alternatives exceeds 7 +/- 2. If you have a lot of alternatives, you can use AHP for the criteria weights only, and combine with another method for alternative evaluation, for example it could be a simple rating table.
Perfectly clear, thanks for your point. I saw many AHP projects where there are some confusion about the real aim of the method and how to use it and that leaded to misleading results. I think that AHP is a flexible option to make your choise easier, but in some case is used in a very rigid way and make things really complicated. So thanks again for the useful example of the video and for your comment. Best regards.
very useful , thank you so much
exactly what i needed
finally i knew what is gona on in this methodology!
Hi professor. Firstly, thank you very much for the video. I have a question. Should we calculate consistency for each sub-criteria?
Consistency is calculated for each set of pairwise comparisons (main criteria + subcriteria)
@@BPMSG Thank you very much for the answer, professor.
Great explanation, thank you
Alternative evaluation can be done using pairwise comparisons (decision matrix) or using any other methods,
Likert scale can not be used for the pair wise comparisons. You need to use the rational AHP scale 9 ... 1 ... 1/9
+Klaus G
Can i use scale >>> 1 2 3 4 5 for questionnaire ?
+Kanyarat Phudphad
You could map 1 2 3 4 5 to the AHP scale 1 3 5 7 9. Important is that your questions are pairwise comparisons, i.e. "what is more important, A or B, on a scale 1 to 9 (in your case 1 to 5)
Thank you so much.
can you recommend AHP book?
+Kanyarat Phudphad
One of the latest from Saaty is Mathematical Principals of Decision Making. Its available from Amazon.
Thank you so much
Where is the Excel sheet?
bpmsg.com/new-ahp-excel-template-with-multiple-inputs/
A good review about AHP can be found in "Review of the main developments in the Analytic Hierarchy Process" from Alessio Ishizaka and Ashraf Labib
so great!
and help me to learn~ Thx a lot!
Goodmorning! BPMSG Can you help me. Do you have a sample program for AHP (Analytical Hierarchy Process) that can help for my study. Thank you in Adnvance.
Sharmaine Garcia Go to my website bpmsg.com. There you can find an online software and a lot of information about AHP
BPMSG Sir Can I get the source code? To know the implementation of the Algo.
The excel template can be unprotected w/o password. You can see all formula. My AHP online software is not open source, but all mathematical relationships are shown in the manual (you can download from my website)
Regards, Klaus
Thank you po Sir! :)
Thanks for this
Awesome! thanks a bunch
Thank you so much sir
Thank's a lot . learn something new
Hat tip to Mr. klaus for erudite lecture on AHP
Thanks helpful video
Still don't get it
very nice
I don't know a specific address to find something abt AHP and marketing. Just google. There is one free online journal on AHP. IJAHP.org. Maybe you find something there.
Thank you very much
Praise Klaus, thank you.
Thank you!
Thank you sir
nice! thank you for this :)
A new AHP excel template with multiple input is now available on mye website bpmsg.com
thank you
The latest update of my AHP excel template with improved accuracy is now available for download from my website bpmsg dot com
thanks master
thank you very much sir! :)
thanks, now I've understand :)
Wow, a 64 MB smartphone. Imagine that.
Show!!
Get the latest information, templates and updates from my blog bpmsg .com
4:55
Check out my free tools on bpmsg.com
Thank you for a nice explanation
Big Thank you. very useful.
Thank you so much