3
$\begingroup$

What to do in these cases where I can't rank by the VIKOR multicriteria method because Q has no value?

    library(MCDM)
    
    df1<-structure(c(14509.9423478068, 4648.5779693212, 922, 876), .Dim = c(2L, 
    2L), .Dimnames = list(NULL, c("Criteria1", "Criteria2")))

> df1
     Criteria1 Criteria2
[1,] 14509.942       922
[2,]  4648.578       876
    
      w <- c(0.5,0.5)
      cb <- c('min','max')
      v <- 0.5
      result1<-VIKOR(df1,w,cb,v)

> result1
  Alternatives   S   R   Q Ranking
1            1 0.5 0.5 NaN       -
2            2 0.5 0.5 NaN       -
$\endgroup$
3
  • $\begingroup$ Related but not a dupe: Using VIKOR multicriteria method in R $\endgroup$ Mar 16 at 22:36
  • $\begingroup$ It is a different question @SecretAgentMan $\endgroup$
    – Antonio
    Mar 16 at 23:19
  • $\begingroup$ Yes, which is why I said it is not a duplicate. My comment is intended to help future visitors find a related question. They may have similar questions as you. I hope they upvote both posts. $\endgroup$ Mar 17 at 12:58

1 Answer 1

2
$\begingroup$

It means that both options have the same ranking and it is up to you what to do with that information.

If you look at the VIKOR paper of Opricovic & Tzeng (2004), you can see that $Q$ is defined as follows:

$$Q_j = v(S_j-S^*)/(S^--S^*) + (1-v)(R_j-R^*)/(R^--R^*)$$ where $$S^* = \min_j S_j, \quad S^-=\max_j S_j$$ $$R^* = \min_j R_j, \quad R^-=\max_j R_j$$

In your example, $S_1$ and $S_2$ are both $0.5$, therefore, $S^*=0.5$ and $S^-=0.5$ which leads to a division by zero and consequently, $Q$ is not defined. This holds not only for $S$ but also for $R$.

Opricovic, S., & Tzeng, G.-H. (2004). Compromise solution by MCDM methods: A comparative analysis of VIKOR and TOPSIS. European Journal of Operational Research, 156(2), 445–455. doi:10.1016/s0377-2217(03)00020-1

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.