A two-phased gradient technique for budget allocation: An application to data collection budget allocation in efficiency measurement of DEA

This paper presents a gradient based algorithm to solve the data collection budget allocation problem in efficiency measurement of DEA. A hard limit is put on how many total readings can be collected and we aim to determine how these should be allocated among the different attributes. A two-phased gradient technique is developed to solve this problem. The numerical results showed that the proposed gradient based algorithm performs well in solving the problem. The insight from the results reveal that, it is important to allocate the budget intelligently and the proposed algorithm serves as a potential method in this area.