Earthmoving productivity estimation using linear regression techniques

The problem of accurate estimation of earthmoving productivity has intrigued many researchers for many years, but there is yet no model that will predict the output of such operations with a satisfactory degree of confidence for all situations. The problem is exacerbated by the occurrence of hauler bunching that results in many deterministic models overestimating productivity. Systems analysis methods can be used, such as queuing theory or simulation, but they are generally impractical to the average site planning engineer. This paper investigates the results obtained from over 140 separate earthmoving operations taken from four different highway construction projects. Initially, the effects of bunching are determined, and many factors that influence productivity are identified. To investigate the relationships between earthmoving operating conditions and both productivity and bunching, and to develop a deterministic model that will enable earthmoving operations to be planned for many different situations with relative ease, stepwise multiple regression techniques are used. The results indicate that there is a strong linear relationship between operating conditions and productivity.