Applying Dependency Structure Matrix and Monte Carlo Simulation to Predict Change in Construction Project

APPLYING DEPENDENCY STRUCTURE MATRIX AND MONTE CARLO SIMULATION TO PREDICT CHANGE IN CONSTRUCTION PROJECT

ZHEN-YU ZHAO, QIAN-LEI LV, WEI-YANG YOU

School of Business Administration, North China Electric Power Univ., Beijing 102206,China

E-MAIL: zhaozhenyuxm@263.net, seawolf47@163.com, ywy132639@163.com

Abstract:

Change management is an important aspect in construction project management, as changes constitute the main cause of schedule delay, cost overrun and other distributions. Many change management systems has been developed to minimize the impacts of change or facilitate changes management. However, change should be predicted as early as possible, and the problems can be identified and beneficial modification can be made as soon as possible. In this paper, an change prediction system using activity-based Dependency Structure Matrix (DSM) was developed to facilitate managing change, which can identify the possible changes in construction project. The outputs of the system for an application have represented the main function of the system. Methods about how to acquire date required for proposed system are also presented. The proposed system can be helpful for project managers to manage change more efficiently.

Keywords:

Change prediction; project management; Dependency structure matrix; Monte Carlo Simulation

1. Introduction

Changes are very common issues in construction project especially in construct process. Any additions, deletions, or other revision to project goals and scope are considered to be changes, whether they increase or decrease the project cost or schedule [4]. In a complex construct process, the cause for change is more complicate. Most commonly, lack of timely and effective communication, lack of integration, uncertainty, a changing environment, and increasing project complexity are the drivers of project change [17]. Because of the cause's complexity, the management of change in construct process is very difficult. However, changes have obvious impacts on construct process, not only on project's schedule and cost but also on project's performance, such as labor efficiency [1]. Furthermore, change is inevitable in construction project sometimes, primarily because of the uniqueness of each project and the limited resources of time and money that can be spent on planning, executing, and delivering the project [2]. Therefore, project management teams must have the ability to respond to change effectively in order to minimize the impact to the project.

1.1. Related research

The most effective method to take change in control is to design efficient change management system in project management system. Some of them have been the component of IT system. IT based methodologies such as computer aided engineering, engineering document management system are used to facilitate managing changes. C. Soh et al proposed a parametric technique based constraint methodology to coordinate design consistency to facilitate managing changes [5]. Furthermore, a reporting system was used to view the history of all changes made by all disciplines. C. Charoenngam had developed a Web-based change order management system that supports documentation practice, communication and integration between different team members in the change order workflow [6]. Analyzing the impact of changes or change orders also have been the mainly function of change management system. Dynamic Planning and Control Method (DPM) had been introduced to simulate the impacts of changes and other conflicts on construction project's performance to facilitate management of changes [14]. However, these methods are mainly focus on how to minimize the change's or change order's impacts and how to facilitate the management process for changes. Kartam has suggested that conflict will be minimized when a problem has been studied as early as possible, since the problems can be identified and beneficial changes can be made [10]. Thus, change prediction theory had been introduced to facilitate managing changes in construction project. I.A. Motawa proposed a fuzzy logic based dynamic planning system for construction project and then the risk of change in the form of "project's stability" can be generated by the system [9]. However, project lever change prediction is not enough for pre-management of changes, detailed information such as change probability and change scope is needed. In this paper, the causes of change in construction project are analyzed in the form of information flow, by using a matrix based methodology. And then, an activity lever change prediction system is designed to simulate the process, during which change happens, to generate activity lever information about changes may happened in construction process.

1.2. Research methodology

The one way to predict changes during Construction Process (CP) is to understand their cause and the process during which change happens, and then use efficient method to simulate this process. In this paper, the cause of change will be analyzed in the form of information flow, and then a method which can graphically represent the information flow in complex system called Dependency Structure Matrix (DSM) is used to represent and analyze the information flow graphically. In addition, rework scope is used to evaluate the effect of the information in the form of rework scope. Thus, the process during which change happens can be simulated by simulating the rework scope's variety. Additionally, Monte Carlo Simulation (MCS) is integrated in the system to enable statistic analysis for the outputs of the system.

2. Dependency structure matrix

In proposed system, DSM is mainly used to model the causes of changes in CP. DSM also known as Design Structure Matrix is a generic matrix-based framework for information flow analysis and is a useful tool for analyzing highly complicated activities and coupled tasks [3]. Despite the common nomenclature, using DSMs for these different purposes requires different methods and objectives. There are mainly four types of DSM had been developed for these purposes [16]. The four types are: (1) component-based or architecture DSM; (2) team-based or organization DSM; (3) activity-based or schedule DSM; and (4) parameter-based or (low level) schedule DSM. In this paper, Activity-based DSM is used to model the process during which change happens in CP.

A basic Activity-based DSM consists of a square