Fuzzy clustering model for estimating haulers' travel time

This paper presents a two-step fuzzy clustering method for estimating haulers' travel time. The proposed method provides a generic tool that can be incorporated in models dedicated for estimating earthmoving production. The estimated travel time takes into account the acceleration and deceleration in the transition zones. The developed method utilizes linear regression and fuzzy subtractive clustering. Seven factors influencing haulers' travel time were first identified and their significance was then quantified using linear regression. The regression analysis was performed utilizing 180 training cases, generated using commercially available software for different models of haulers. The data were generated randomly to represent a wide range of possible combinations of factors affecting travel time of haulers across different types of road segments. The training data were subsequently used in the development of the proposed method. Unoptimized subtractive clustering, optimized Takagi-Sugeno zeroth-order subtractive clustering, and optimized Takagi-Sugeno first-order subtractive clustering were used in estimating haulers' travel time. Their performance was evaluated using 36 test cases, also generated randomly in a similar manner to those utilized for training. The optimized Takagi-Sugeno first-order subtractive clustering model was found to outperform the other two, and was accordingly used in the proposed method. A numerical example is presented to demonstrate the use of the developed method and illustrate its accuracy.