Page 34 - Azerbaijan State University of Economics
P. 34

3.   Fuad Selamzade: Measurement of the Efficiency of Azerbaijan Region Hotel
                             4.                                  Organization:  Window Analysis and Malmquist Index



                                         virtual outputs  =     1    1    +    2    2    + ...+                           (1)
                                          virtual inputs     1    1    +    2    2    + ...+              

                    Where; us – s. expresses the weight of output, yso – “o” expresses s. output obtained
                    from “o” unit, vm – m. weight of input, xmo - “o” expresses m. input used by “o” unit .

                    Supposing that a DMU produces outputs yr, r=1, 2, …, s from inputs xi, i=1,2,…,m,
                    equation can be written as follows with the help of the appropriate weights on the
                    variables (ur=1,2,…,s; vi=1,2,…,m) (Charnes et al., 1978: 431):

                                                   =  ∑      =1                                    (2)
                                                     ∑                  
                                                        =1

                    As  it  can  be  seen  from  the  formula,  DEA  should  be  considered  as  a  conceptual
                    model  in  a  sense,  because  the  fractional  program  uses  the  ratio  of  total  factor
                    efficiency.

                    The  linear  program  for  DMUo  is  made  by  equalizing  the  denominator  of  the
                    objective function in the fractional function to 1 (Charnes et al., 1978: 431).

                                              = ∑             ;       ∑           = 1                                     (3)
                                                           
                                                                           
                                                                  =1
                                                 =1

                                      ∑      =1                        
                      Constraints:              ≤ 1 (   = 1, 2, …   ); ∑        = 1;    > 0 ;     > 0
                                       ∑                               =1                      
                                          =1

                    Equation (3) is a linear equation and restricts the weighted sum of inputs to 1 and
                    maximizes the weighted output sum of “o” DMU by selecting appropriate values for
                    ur vi. At the same time, the efficiency value cannot exceed 1.

                    When Banker, Charnes and Cooper created the BCC model, they defined the set of
                    production possibilities as follows (Cooper et al., 2011: 88);

                                                PB={(x, y)| x ≥ Xλ, y ≥ Yλ,  eλ=1, λ≥0 }           (4)

                                          sxn
                                                 n
                    X=(xj)εR mxn , Y=(yj)εR , λεR  e is a sequence vector whose all elements equal to 1.
                    According to the above definition, the only difference that distinguishes the BCC
                    model from the CCR model is that the constraint is added to the model.

                                                  λ = ∑     λ = 1                                  (5)
                                                              
                                                         =1

                    This constraint, together with the condition λj≥0, enables the realization of different
                    combinations of n piece of DMU only within a concave efficiency upper boundary
                    line (Banker et.al, 1984: 1086).


                                                           34
   29   30   31   32   33   34   35   36   37   38   39