Estimating excavation duration: OOP plus NN approach

Abstract

This paper presents a computing method based on object-oriented programming (OOP) and neural networks (NNs) for determining the minimum duration of an excavation task. The concept of optimizing the task plan for an excavation is explained first. In devising an object-oriented solution to the defined problem the needed classes are defined in terms of the attributes and functions of objects for the task, the task plans, and the excavator involved. A trained neural network is used in the program to evaluate the cycle times of the employed excavator performing in given job conditions. The approach is illustrated by an example involving an excavator digging and repositioning for a basement excavation. The example demonstrates that the object-oriented method with a neural network representing the excavator offers a highly modular and flexible way of assessing excavation task duration in a variety of scenarios.