Client-Server Interaction Knowledge Discovery for Operations-level Construction Simulation Using Process Data

Abstract

For the past few decades, construction researchers have investigated the potential of simulation modeling in providing decision-makers with insights into different aspects of a project. Simulation can assist in studying the performance of an engineering system during early planning, pre-construction, execution, and maintenance. This is typically done by providing means and methods for preparing work plans, look-ahead schedules, and what-if analysis. However, most existing construction simulation paradigms address long-term planning and scheduling needs. Work is still needed to disseminate the use of simulationbased decision-making in short-term operations-level planning and control during project execution phase. Simulation models, and in particular, discrete event simulation (DES) models are often employed to get insight into the performance of systems that are repetitive yet stochastic in nature. A good example of such systems that can be found in almost all industries is queuing systems. In construction, queues are ubiquitous due to specific project plans, resource allocation patterns, or operational bottlenecks. Certain client-server interaction schemes determine the properties by which a queue is characterized and subsequently modeled in a computer simulation model. This paper presents a methodology that enables data collection, fusion, and mining from construction resources in a queuing system to discover necessary knowledge and generate and update corresponding simulation models.