مکانیسم مدل سازی یکپارچه برای مدل بهینه شبیه سازی عملیات های ساخت و ساز
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|9216||2006||14 صفحه PDF||سفارش دهید||6250 کلمه|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : Automation in Construction, Volume 15, Issue 3, May 2006, Pages 327–340
A planner may use the discrete-event simulation to analyze and design the construction operation process that optimizes the overall performance of a construction system. Normally, the basic elements used in construction operation process simulation system, such as CYCLONE (CYCLic Operation NEtworks), are “activity” and “queue.” Activity is used to model the task which consumes resources and takes time to perform. Queue acts as a storage location for resources entering an idle state. In the simulation system, queues have to be created according to the ways of assigning resources to activities. Conventionally, planner defines queues according to his/her judgment by determining which and what amount of resources should be allocated to which activity. Consequently, various modeling schemes have to be examined to obtain the best simulation model. However, such a process of creating queues and activities is time consuming and requires iterations. This paper introduces a Genetic Algorithms (GA)-based modeling mechanism to automate the process of selecting the optimal modeling scheme. Case study shows that this new modeling mechanism along with the implemented computer program not only can ease the process of developing the optimal resource combination but also improve the system performance of the simulation model.
Discrete-event simulation has been used to assist construction engineers in analyzing and designing construction operation processes for years. One of the advantages to utilize simulation in designing construction processes is that planners may examine various schemes of the simulation model to better understand how resources influence the overall performance of a construction system so as to select a better resource assignment. Usually, there are two ways to conduct the above experiment. One is to establish a single model and then conduct simulations by examining different resource combinations to find out which resource combination optimizes the system performance. Another, a more complicated one, is to test each possible scheme of model that involves in building various simulation components within the simulation model and then test the system performance through running all resource combinations in each modeling scheme. The former only has to deal with the available resource combinations; however, the latter one has to deal with the modeling schemes and the resource combinations. The traditional way to determine the optimal resource combination of a single simulation model is to exhaustively examine all resource combinations. Riggs  proposed a computer-based program called sensitivity analysis for facilitating such enumeration. However, if possible resource combinations increase explosively, sensitivity analysis could be extremely time consuming. As a result, AbouRizk and Shi  proposed a heuristic algorithm (HA) which efficiently locates the most appropriate resource allocation of the simulation system. Whereas, the solution of their heuristic approach is usually the local optimum and the performance of such an approach is problem-dependent. Therefore, Cheng and Feng  integrated Genetic Algorithms (GA) and simulation to efficiently find the optimal resource combinations in terms of different objectives, such as minimizing the unit cost or maximizing the productivity rate of the simulation model. Similarly, Hegazy and Kassab  used GA simulation technique for resource optimization in construction planning. Furthermore, Cheng et al.  combined HA and GA to improve the efficiency of only using GA for locating the optimal resource combination. Though several approaches were proposed to find the optimal resource combination of a single simulation model, determining the optimal scheme of a simulation model which involves building various simulation components within the simulation model and finding the optimal resource combination have not been well explored. This paper proposes a modeling mechanism that uses Genetic Algorithms (GA) as the optimization engine to automate the process of selecting the optimal modeling scheme of a simulation model. In addition, a computer program that integrates the proposed with CYCLONE (CYCLic Operation NEtworks) simulation methodology is also presented in this paper.
نتیجه گیری انگلیسی
Simulation technique is easily applied to modeling construction operations. Conventionally, the construction planner has to create queue elements for storing resources used in activities. If there are more than one modeling scheme of developing the simulation model, the construction planner has to test each scheme of the simulation model to find the suitable one. Moreover, if there is large number of resource combinations that should be examined in each modeling scheme, it becomes uneconomical to examine every combination in all models. This research proposes a modeling mechanism that allows the construction planner to focus on modeling the construction processes rather than examining all possible schemes of a simulation model that has the best resource utilization. From the case study, this new modeling mechanism along with the implemented computer program not only can ease the process of developing the optimal resource combination but also improve the system performance of the simulation model.