Learning construction-planning knowledge from experience

Abstract

The evolution of a construction product throughout ifs engineering life cycle requires many experts in different fields. A training process for a novice to become an expert requires tremendous cost and time. When an expert leaves or retires, knowledge that is accumulated through years of experience is also gone, or stored as various documents that are unorganized and hard to find. Human planners are one type of experts that require such a learning process, which is slow and inefficient. This paper uses construction-planning knowledge as an example to describe an knowledge acquisition system that extracts planning knowledge through generalization of existing construction schedules with human annotation.