“…The specific dual‐failure restorability is, therefore, defined by the resulting working portion of the network after the occurrence of that specific dual failure. This is denoted by R 2 ( i , j ) and it can be measured by (12) , where N ( i , j ) is the NWC due to the dual failure occurring in nodes i and j ; w i is the number of working units on span i , and w j is the number of working units on span j . If we consider all the possible dual‐failure scenarios and compare those with the resulting working units for all these cases, as done in the work by , then the network‐wide average dual‐failure restorability, denoted by R 2 , can be calculated by (13).…”