WEBVTT Kind: captions Language: en-IN 00:00:17.040 --> 00:00:28.880 okay hello everyone we have been discussing the topic on supplier selection supplier evaluation so 00:00:28.880 --> 00:00:37.120 always you know such kind of decisions they invite lot many factors to be considered and so that you 00:00:37.120 --> 00:00:46.040 know one can arrive to some index or some score with the help of what we can rank the suppliers 00:00:46.040 --> 00:00:55.280 we can then select the best suppliers okay so to handle such kind of situations sometimes 00:00:55.280 --> 00:01:03.120 you know we need some methods which are classified under multicriteria decision making 00:01:03.120 --> 00:01:09.800 models there are many multicriteria decision- making model exist right every model has its 00:01:09.800 --> 00:01:18.560 own advantages and disadvantages okay so one of them is called topsis This is technique 00:01:18.560 --> 00:01:28.400 for order preference by similarity to the ideal situation or ideal solution okay so what it does 00:01:28.400 --> 00:01:37.880 it it can accom dat multiple factors which are also called attributes and criteria in the last 00:01:37.880 --> 00:01:45.080 session we have seen while selecting suppliers we have to have many attributes or criteria on 00:01:45.080 --> 00:01:54.520 that basis one has to select or evaluate the suppliers okay so how we can arrive to a composite 00:01:54.520 --> 00:02:01.360 score because there are so many criterias and at the same time we have many suppliers existing you 00:02:01.360 --> 00:02:10.560 know who who are eligible who are qualified to supply the products so which one is the best one 00:02:10.560 --> 00:02:20.680 when one has to evaluate or you know when we have to select for many criteria or many factors okay 00:02:20.680 --> 00:02:28.640 so with this backdrop we will go through this tool as I've said this is a multi criteria decision 00:02:28.640 --> 00:02:37.560 making tool correct which fits very well in the problem domain of supplier selection the objective 00:02:37.560 --> 00:02:47.200 is to assess and rank various Alternatives and in our context the alternatives are nothing but 00:02:47.200 --> 00:02:55.320 suppliers number of suppliers right which are based on certain attributes now attributes in 00:02:55.320 --> 00:03:00.880 the last class we have seen you know it can be management capability it can be quality related 00:03:00.880 --> 00:03:08.960 attribute it can be delivery related attribute right so the set of attributes represent the 00:03:08.960 --> 00:03:15.440 factors that influence the decision Maker's Choice while prioritizing the Alternatives 00:03:15.440 --> 00:03:24.400 so this is the aim of the topsis right it will help us in giving relative significance because 00:03:24.400 --> 00:03:32.880 each attribute is rated differently by different organization right or it can be rated differently 00:03:32.880 --> 00:03:41.040 by different decision makers okay so this method prioritizes the alternative the concept they use 00:03:41.040 --> 00:03:49.360 is the ukian distance okay so we will see how they use this ukian distance right so the concept of 00:03:49.360 --> 00:03:57.680 topsis is this prioritize the Alternatives that is prioritize or rank the suppliers according to 00:03:57.680 --> 00:04:05.080 ukian distance so the moment we say distance me means like you know if there is some ideal 00:04:05.080 --> 00:04:12.640 situation or ideal solution okay if there's a positive idle solution means like you know if 00:04:12.640 --> 00:04:22.960 any supplier qualifies based on the criteria if any supplier qualifies to be as close as possible 00:04:22.960 --> 00:04:33.520 to the positive solution okay so that would be the best solution at the same time the longest 00:04:33.520 --> 00:04:42.080 distance from the negative ideal solution okay so we based on the two indicators we 00:04:42.080 --> 00:04:49.400 would bring an overall index okay which will be determined using distances from idle Solutions 00:04:49.400 --> 00:04:57.000 okay so what are the inputs required obviously we need some Alternatives that is Alternatives 00:04:57.000 --> 00:05:08.000 in our context can be suppliers so what are the Alternatives S1 S2 S3 S4 can be the suppliers 00:05:08.000 --> 00:05:14.920 this is the Alternatives one can have to choose from or rank them right and at the 00:05:14.920 --> 00:05:24.040 same time we have attributes right this is our a and this is the attributes so attributes can 00:05:24.040 --> 00:05:35.560 be quality it can be delivery and so on on right so attributes will be or factors will be 00:05:35.560 --> 00:05:50.160 will okay so for each alternative we would be having some scores for the attribute okay so 00:05:50.160 --> 00:05:56.520 this is the decision makers like you know we'll be having different scores right if scoring scheme is 00:05:56.520 --> 00:06:04.440 one to five right one supplier can get from for Factor one maybe five right another supplier will 00:06:04.440 --> 00:06:13.640 get you know two right if it is scoring system is between 1 to 10 so accordingly one one can consult 00:06:13.640 --> 00:06:24.480 the experts right in that area so to rank them or to score them right so we need Alternatives 00:06:24.480 --> 00:06:32.000 from which we need to select we need attributes or criteria on on what bases we are going to select 00:06:32.000 --> 00:06:41.400 so this are the scores assigned by decision makers in The Matrix right so what are the topsis method 00:06:41.400 --> 00:06:52.000 steps we would use right remember we need idle solution right and we need to find out two 00:06:52.000 --> 00:06:59.720 idal solutions right one is best one is another is worst right so that we can see who are the 00:06:59.720 --> 00:07:08.000 suppliers closer to best and farest from worst this is the concept okay this is the concept 00:07:08.000 --> 00:07:17.240 of the topsis method now to do that what we can do is so we will first take the decision makers which 00:07:17.240 --> 00:07:25.400 assign the scores and create a decision Matrix like this is the decision Matrix once we get the 00:07:25.400 --> 00:07:34.280 score now after getting the scores one an object that you know probably somebody's giving score 00:07:34.280 --> 00:07:44.720 of seven here right it might be like you know the there is a subjectivity like you 00:07:44.720 --> 00:07:51.760 know in providing scores like here somebody might give nine right so there can be a subjectivity of 00:07:51.760 --> 00:08:01.480 scores right or one can question like you know how this scores can be taken as it is right right 00:08:01.480 --> 00:08:12.680 unless we have something to normalize the Scopes right so to cater to the subjectivity right in 00:08:12.680 --> 00:08:19.600 this because this there's a lot of subjective judgment goes like you know we get the experts 00:08:19.600 --> 00:08:28.040 opinion right in in terms of providing the scores okay so it is very important first to normalize so 00:08:28.040 --> 00:08:36.760 we will get the normalized performance ratings out of this decision Matrix and then we will 00:08:36.760 --> 00:08:47.360 get the normalized decision Matrix now how to normalize for each attribute right we can get the 00:08:47.360 --> 00:08:56.160 score right okay when when we will go to the step with the example it would be more 00:08:56.160 --> 00:09:03.720 clear right so here the mathod meod we would use is that you know we will take the score we will 00:09:03.720 --> 00:09:12.480 divide the score with the summation of square of the scores given to the all attributes okay so we 00:09:12.480 --> 00:09:19.280 will see that as we will proceed with the example the idea is first to normalize the scores okay 00:09:19.280 --> 00:09:26.640 when we would normalize the scores it is very important to have the weighted scores because 00:09:26.640 --> 00:09:34.880 for any organization all the attri rutes might not carry equal priority okay they might have 00:09:34.880 --> 00:09:41.840 different priority so based on the priority we can assign weights so that we can get a weighted 00:09:41.840 --> 00:09:50.440 normalized Force then we will find out something to get the positive ideal solution and negative 00:09:50.440 --> 00:09:58.080 ideal solution so what are the indicators of positive ideal solution so that we can compare 00:09:58.080 --> 00:10:06.160 every possibility compared to the positive idal solution at the same time we need to calculate 00:10:06.160 --> 00:10:13.920 the every possibility with the negative ideal solution okay so that would be done then we will 00:10:13.920 --> 00:10:22.560 get the separation value right so we will see that you know which alternative has done well in terms 00:10:22.560 --> 00:10:30.920 of meeting this conceptual understanding of you know topsis which supplier has been you know more 00:10:30.920 --> 00:10:39.000 closer to positive ideal and farther to the negative ideal okay so the separation value will 00:10:39.000 --> 00:10:45.880 let us tell us this and then we will determine the overall performance go okay so these are 00:10:45.880 --> 00:10:55.280 the steps we would follow keep in mind all the steps step one let's do this with an example if 00:10:55.280 --> 00:11:04.880 we are required to assess or prioritize three suppliers that is three Alternatives so here 00:11:04.880 --> 00:11:12.040 in our context this is a supply selection problem or Supply ranking problem we want to assess three 00:11:12.040 --> 00:11:19.200 Alternatives right so what are the Alternatives supplier one two and three and let's say we have 00:11:19.200 --> 00:11:27.360 three attributes reputation trust and ontime delivery right now this multicriteria decision 00:11:27.360 --> 00:11:35.200 making models they also give us one liberty right many of times there are indicators like you know 00:11:35.200 --> 00:11:41.080 which cannot be measured with the numerical numbers like you know we cannot measure them 00:11:41.080 --> 00:11:49.040 objectively okay for example your delivery time you can measure right when the delivery 00:11:49.040 --> 00:11:55.400 come at what time how many hours it takes or how many minutes it takes we can quantify it you know 00:11:55.400 --> 00:12:02.160 we can use the exact number to arrive to some SCH SC right but there are many indicators like 00:12:02.160 --> 00:12:09.120 trust right whether the supplier is trustworthy whether the supplier is dependable so such kind 00:12:09.120 --> 00:12:17.040 of indicators are very difficult to quantify and you know assign a number so in such cases 00:12:17.040 --> 00:12:25.560 one can take opinions of expert that okay on the scale of 1 to 10 where you rank them or what 00:12:25.560 --> 00:12:31.920 score you would like to give right based on maybe past experience or based on the past performance 00:12:31.920 --> 00:12:41.280 of the suppliers so those indicators can also be accommodated in mcdm methods and so is in 00:12:41.280 --> 00:12:49.680 toses so we will get the scores right suppose so between 1 to 10 you know the experts have given 00:12:49.680 --> 00:12:57.840 the scores so we have got the scores but we need to normalize right so for normalize second step 00:12:57.840 --> 00:13:07.600 we need to normalize it so how we are going to normalize the score right divided by for 00:13:07.600 --> 00:13:14.480 this attribute what are the other scores okay so that would give like you know because this 00:13:14.480 --> 00:13:23.600 is this can be like given keeping the relative standing of suppliers okay so such 00:13:23.600 --> 00:13:30.600 normalization would help us you know to remove any subjectivity right while giving the scores 00:13:30.600 --> 00:13:39.360 to the suppliers so for each cell what we can do we can take that score divided by the squares of 00:13:39.360 --> 00:13:47.880 the score for each attribute take square root and we will get you know the normalized scores 00:13:47.880 --> 00:13:59.040 for each attribute for all the Alternatives okay so now this is a normalized matrix we have have 00:13:59.040 --> 00:14:08.680 got the normalized scores now okay so now we have normalized the data here okay now what is 00:14:08.680 --> 00:14:20.400 the next step next step is whether each attribute carries equal weight whether for any organization 00:14:20.400 --> 00:14:31.640 is it ontime delivery more important then trust or reputation or vice versa okay so in that case 00:14:31.640 --> 00:14:42.680 right I need to consider weights suppose I give weights in this form right so if that is the case 00:14:42.680 --> 00:14:55.160 I know the weights of each attribute now I would get the step three weighted normalized Matrix 00:14:55.160 --> 00:15:07.120 where I will multiply each normaliz I score with the respective weight of the attribute right so 00:15:07.120 --> 00:15:16.320 now I get the weighted normalized scores okay so this is the step three weighted normalized score 00:15:16.320 --> 00:15:24.760 right so now what now our Matrix is ready which is normalized which has captured the importance of 00:15:24.760 --> 00:15:35.000 criteria or importance of attributes okay now now we need to go forward to find the positive ideal 00:15:35.000 --> 00:15:45.120 or negative ideal solution so for that what we will do right we have you know all this for each 00:15:45.120 --> 00:15:53.400 attribute we have the weighted normalized score for each alternative or for each supplier we can 00:15:53.400 --> 00:16:05.200 see now out of this one which one has done the best and which one has done the worst that would 00:16:05.200 --> 00:16:19.320 help so this is U so this is by the way this is my U Vector with you know all the elements as uig 00:16:19.320 --> 00:16:28.200 J so if this is uig J means this is the Matrix U which is weighted normalized Matrix in this I 00:16:28.200 --> 00:16:42.640 would like to find for each attribute which one is the benefit attribute okay benefit attribute 00:16:42.640 --> 00:16:59.240 means which one is the best performer right so that uig J will be the U1 plus okay that means 00:16:59.240 --> 00:17:07.600 for that attribute I will get this one for first attribute and then second so on for all attributes 00:17:07.600 --> 00:17:18.480 I would get okay so and when this cost attribute will come what do we mean by cost attribute any 00:17:18.480 --> 00:17:32.400 idea pardon undesirable undesirable traits are the undesirable traits everyone is with this 00:17:32.400 --> 00:17:40.680 argument that this is the undesirable if this is undesirable then why we have to find the negative 00:17:40.680 --> 00:17:46.920 solution because undesirable thing will be covered 00:17:46.920 --> 00:17:54.240 in the negative solution so what does this mean so that definitely is not 00:17:54.240 --> 00:18:06.560 undesirable exactly less of that attribute is better okay so what happens like you know if 00:18:06.560 --> 00:18:20.040 suppose if suppose in this scor right I would say that the cost attribute right the Lesser 00:18:20.040 --> 00:18:28.880 it is the better it is lesser the better right if it is the cost one so it is other way around if I 00:18:28.880 --> 00:18:36.480 would say that lower the score better means lower the cost is better right in that case toses will 00:18:36.480 --> 00:18:46.680 accommodate that also but in that case rather than getting the positive ideal solution to consider 00:18:46.680 --> 00:18:56.960 Max one must consider Min okay right because you know when you are solving a real business 00:18:56.960 --> 00:19:06.640 problem there there would be some performance indicators magnitude wise or numerical value 00:19:06.640 --> 00:19:15.040 wise some would be more the better some would be lesser the better right like you know your 00:19:15.040 --> 00:19:22.960 cost the lead time lesser the better and if you would score like you know lesser the better okay 00:19:22.960 --> 00:19:34.680 in that case your positive idle solution should be minimum value okay so this is the advantage of 00:19:34.680 --> 00:19:47.120 topsis over many of other multi criteria decision making model okay so this is very important aspect 00:19:47.120 --> 00:19:54.680 here the moment you will determine the positive idal solution remember that right so what is the 00:19:54.680 --> 00:20:02.280 maximum if it is benefit attribute minimum if it is cost attribute and you know you can determine 00:20:02.280 --> 00:20:10.440 the negative idle solution right which should be minimum if it is benefit attribute it should be 00:20:10.440 --> 00:20:19.440 maximum if it is a cost attribute right so I hope you all are with me now here we have to find what 00:20:19.440 --> 00:20:29.440 should be the a plus we need to find out for that I need to get the set what is the set u 1+ I need 00:20:29.440 --> 00:20:41.080 to get U1 plus I need to get U2 plus I need to get u3 + right so how to get U1 plus which is 00:20:41.080 --> 00:20:52.240 the maximum UI J for the attribute so maximum of this one will give me the A+ and this will be U 00:20:52.240 --> 00:21:02.400 1+ okay so I will get U 1+ 1.98 right and in the next is 00:21:02.400 --> 00:21:18.320 2.68 and so on we all are at same page okay so we have got now so now things are ready with 00:21:18.320 --> 00:21:31.840 us right here we have computed positive idle solution and negative idal solution okay 00:21:31.840 --> 00:21:46.120 yes yeah the calculation thing you know you can do that right so this is you will 00:21:46.120 --> 00:21:51.680 get you know the positive and negative ideal solution has been determined right 00:21:51.680 --> 00:21:59.160 now things are intuitively what we have discussed in the beginning now what our 00:21:59.160 --> 00:22:04.080 intuition says we have idal Solutions with us now now what we are supposed to 00:22:04.080 --> 00:22:19.600 do how far we are from ideal Solutions okay so that is easy how far each cell is from the idle 00:22:19.600 --> 00:22:27.240 solution and how far it is from the negative idal solution right so this will be computed 00:22:27.240 --> 00:22:38.760 like that for each cell this minus this square plus this minus this square right so we will 00:22:38.760 --> 00:22:47.080 compute that that would give you give you know for each supplier right what is the separation 00:22:47.080 --> 00:22:58.920 value right using the distance measure will be there okay so this is how we will determine you 00:22:58.920 --> 00:23:09.960 know the difference or distance between the ideal solution and the cell right and then 00:23:09.960 --> 00:23:21.360 what is the index so index will be you know D minus divided by di minus and plus di so this 00:23:21.360 --> 00:23:32.480 is more of a similarity right so we will use this index to consol solidate you know our distances 00:23:32.480 --> 00:23:40.840 from the ideal Solutions okay and then we will find out the similarity right so after Computing 00:23:40.840 --> 00:23:50.680 this one using our data from the separation values we will compute that and so the best 00:23:50.680 --> 00:23:58.640 alternative out of this thing is this one so this is the best alternative that supplier S2 and the 00:23:58.640 --> 00:24:07.680 worst alternative is supplier S1 okay so this is how one can arrive you can we can rank the 00:24:07.680 --> 00:24:17.560 Alternatives H the suppliers okay or we can select the best supplier right among many Alternatives or 00:24:17.560 --> 00:24:24.680 many suppliers okay so this is how we can use this method right this is a small example with 00:24:24.680 --> 00:24:33.560 three suppliers right but in in reality one can expand like you know in multiple supplier 00:24:33.560 --> 00:24:45.320 situation okay this is very much used like you know in other industry also right whenever you 00:24:45.320 --> 00:24:53.920 have any decision making situation to select a product to select suppliers as one but to 00:24:53.920 --> 00:25:00.600 select you know your financial portfolios right or maybe if you want to compare you know some kind of 00:25:00.600 --> 00:25:08.120 credit risk or default risk like you know so you have to predict in advance like you know 00:25:08.120 --> 00:25:14.120 how the situation can go like you know maybe if you want to rate the customers for default risk 00:25:14.120 --> 00:25:21.520 or credit risk you know you can use the similar kind of methods basically where the decision 00:25:21.520 --> 00:25:27.480 making involves many factors right and all factors are important to be considered and 00:25:27.480 --> 00:25:35.000 all factors carry different weightages right so those kind of situations one can use this 00:25:35.000 --> 00:25:44.440 topsis okay so this is all because in the last class we have seen the weighted method to 00:25:44.440 --> 00:25:51.840 weigh the suppliers okay so that method you know we are simply adding the scores right so there in 00:25:51.840 --> 00:25:59.280 that method right the it can accommodate like you know know which one is the best based 00:25:59.280 --> 00:26:07.120 on the weights given right but this method would help us like you know it's not only the maximum 00:26:07.120 --> 00:26:16.960 weight you know how far that alternative is from the negative solution okay so it's not only the 00:26:16.960 --> 00:26:25.160 closeness or the weighted positive solution but it is also important to to capture that factor 00:26:25.160 --> 00:26:31.640 should be far from the negative C ution okay so this is the plus point of having this technique 00:26:31.640 --> 00:26:41.360 over simple weighted weighted average method for selecting these suppliers okay we 00:26:41.360 --> 00:26:50.360 have just rated the suppliers right but you know as mentioned in the last session that still if we 00:26:50.360 --> 00:26:57.400 have to take further Advanced decisions like you know how to allocate similar quantity or 00:26:57.400 --> 00:27:04.560 or or same quantity among multiple suppliers right which one supplier should get how much 00:27:04.560 --> 00:27:11.440 quantity right and if we have to talk about multiple products how one can do that this can 00:27:11.440 --> 00:27:18.680 be covered in the advanced optimization models so that we will cover in the next session okay thank 00:27:18.680 --> 00:27:25.480 [Music] you