According to some definitions, the ratio of output to input is efficiency. In general, there are two main, parametric and non- parametric, approaches for measuring efficiency. Some parametric methods which are based on econometric models are frontier production function both deterministic and stochastic. On the other hand, non- parametric approach based on mathematical programming. Data envelopment analysis (DEA) a non parametric method is used in this study for evaluation of Iranian railway industry. The most important feature of the DEA is its applicability for the firms that have multiple input and outputs.
Using DEA, we study the efficiency of 13 the region of Iranian railway in 1379, assuming both constant and variable return to scale. The results show that, subject to former assumption, efficiency in Tehran, Esfahan, Hormozgan & Khorasan are maximum and the average efficiency is equal to %86, which means, that the idle capacity is 14 percent in that year. On the other hand, with the later assumption, Zahedan and Lorestan were added to the list too, and average efficiency is equal %93.
So, it is subjected that for increasing efficiency of each region, subject to its constraints, it should apply resources like “referenced firm”.