Fuzzy reasoning expert scheduling system (FRESS) for highway construction subject to rain impact: A case study in Taiwan (China).

摘要

In Taiwan, more and more construction projects adhere to calendar-day-based contracts, whereby the construction contractor will be taking more risks than ever before. Construction scheduling contains uncertainties such as conditions of weather and site, and performance of equipment and labor. Among these uncertainties, rainfall is regarded as a major factor that has severe adverse impact on productivity and duration of highway construction projects. The most common approach for schedulers to account for the effect of rain in Taiwan is by adding a percentage of time to activities based on time of year and type of activity. However, this procedure may not be uniformly applied to all the activities. In addition, the precise effect of rain is difficult to assess and evaluate because it is usually expressed qualitatively rather than quantitatively. Moreover, schedulers usually estimate rain influence based on their past experience in a similar situation. However, their experience is often limited to particular project types, locations, and rainfall patterns. Thus, a more accurate and reliable method is necessary to assist construction schedulers.; This study presents a model to estimate the impact of rain on productivity and duration of highway construction, taking into account linguistic variables such as rainfall level, soil drainage, exposure level of an activity, and work or no-work. This model employs the techniques of expert system, fuzzy set and Mamdani's fuzzy reasoning method. An expert system is utilized because the estimates of activity duration are usually made by experienced schedulers. Fuzzy set theory and fuzzy logic are applied because rain effect is usually expressed linguistically. Based on the model, an intelligent system, named FRESS (Fuzzy Reasoning Expert Scheduling System), is developed for estimating rain influence before and during construction for scheduling, and after construction for claim analysis of delay. FRESS interfaces with the Kappa-PC expert system shell and MS-Project scheduling package, which provides a user-friendly interface. This system contains a knowledge base and database. The knowledge base consists of if-then production rules, obtained from experts in highway construction. The database includes historical daily rainfall data over the past 20 years for nine major cities across Taiwan. A real-world highway construction project is utilized to illustrate the essential features of the system and demonstrate its performance and effectiveness, which can provide construction schedulers accurate and reasonable schedules.