Knowledge-based schedule generation and evaluation

The straightforward and effective realization of construction projects is heavily dependent upon the planning of the construction process and the response to project modifications during the construction phase. Construction-process planning is one of the most important requirements for a satisfactory result, but it is often not given sufficient consideration. The construction planning usually involves great amount of process steps and is developed by the processing of many project participants. The construction-planning steps range from planning of building design and creation of construction drawings up to the generation of schedules for the realization of the planned building. These steps are often executed without overall coordination and reiteratively. The planning process becomes error prone and thus results in inaccurate schedules. The planned construction process must then be operational adapted at the construction site. These drawbacks are mainly produced during the manual creation of schedules that is based on no current or insufficient information. Moreover, due to the complexity of the planning task in the domain of construction planning the manually generated schedules are often imprecise. Furthermore, schedules are not updated when project conditions change. In such situations, project managers balance the available possibilities and intuitively choose one construction alternative for the realization. During this process many factors are often taken into account insufficiently or not at all. In addition, the project managers are wasting their resources generating new schedules for recurring situations that are similar to already executed projects. In consequence, the project experience becomes personal and is not available for other project participants. To eliminate these drawbacks, a knowledge-based approach for the automatic generation of schedules and their evaluation is presented in this paper using the building information for identification of the scheduling subject. With the support of the system, construction schedules can moreover be re-generated flexibly, each time a project is modified, considering the current execution alternatives. All learned experience is stored within a reasoning system for further reuse. Using a decision-support system, construction alternatives can be evaluated considering all relevant factors and the best alternative for the realization can be proposed.