Construction risk assessment is the final and decisive stage of risk analysis. When highly changeable conditions of works execution are predicted, risk should be evaluated in the favorable, moderate, and difficult random conditions of construction. Given the random conditions, the schedule and cost estimate of the construction are developed. Based on these values, the risk of final deadline delay and the risk of total cost increase of construction completion are calculated. Next, the charts of the risks are elaborated. Risk changes are shown in the charts and are analyzed in the range [1, 0].
The basic element of a project organizing construction works is a schedule. The preparation of the data necessary to specify the timings of the construction completion as indicated in the schedule involves information that is uncertain and hard to quantify. The article presents the methods of building a schedule which includes a fuzzy amount of labour, time standards and number of workers. The proposed procedure allows determining the real deadline for project completion, taking into account variable factors affecting the duration of the individual works.
Redundancy based methods are proactive scheduling methods for solving the Project Scheduling Problem (PSP) with non-deterministic activities duration. The fundamental strategy of these methods is to estimate the activities duration by adding extra time to the original duration. The extra time allows to consider the risks that may affect the activities durations and to reduce the number of adjustments to the baseline generated for the project. In this article, four methods based on redundancies were proposed and compared from two robustness indicators. These indicators were calculated after running a simulation process. On the other hand, linear programming was applied as the solution technique to generate the baselines of 480 projects analyzed. Finally, the results obtained allowed to identify the most adequate method to solve the PSP with probabilistic activity duration and generate robust baselines.