Multi-objective optimization algorithms for the activity sequencing problem in steel construction projects

With the current real estate market growth, the productive resources tend to become scarce and expensive in civil construction. Due to this fact, the best use of productive resources has become extremely important for the success of this type of enterprise. Another important fact that has been observed is the increasing use of steel in civil construction, replacing conventional materials like concrete. This fact is due to the aesthetic and quality advantages that this type of construction has been showing in various types of projects. However, besides these advantages provided by the use of steel construction systems, the reduction of the duration and cost of construction and the increased productivity are key factors for its success. However, to achieve these factors, the works need to be much more controlled, which means projects better elaborate where technology is a differential for companies that invest in it. The lack of planning and/or bad planning and guidance on project management have been the main responsible for problems that happen on civil construction. Correct project management is capable of providing reduction of duration and costs, better utilization of productive resource, minimization of risks, and error reduction on the production process. Several tools can be used by engineering aiding the decision making related to project management, within which optimization is emphasized; this is seldom applied in civil construction. Several optimization problems related to projects, which has a wide diversity of real applications, can be found in literature. An important example is the resource-constrained project scheduling problem with precedence relation (RCPSPRP), considering that the correct project activities sequencing results in a better use of the available resources, and consequently, a gain in productivity and time. In this work the RCPSPRP is addressed as a multi-objective optimization problem and aims at minimizing two criteria: the makespan and the total weighted start time of the activities. To solve the problem, five multi-objective algorithms are analyzed, based on Multi-objective GRASP (MOG), Multi-objective Variable Neighborhood Search (MOVNS) and Pareto Iterated Local Search (PILS) methods. The proposed algorithms use strategies based on the concept of Pareto Dominance to search for solutions and determine the set of non-dominated solutions close to the Paretooptimal front, allowing the project manager to choose a solution that fulfills his interests by making the project better planned and controlled. The set of non-dominated solutions attained by the algorithms for a set of instances adapted from literature are compared using four multi-objective performance measures: distance metrics, hypervolume indicator, epsilon metric and error ratio. Statistical experiments were also conducted to prove the existence of a significant difference between the proposed algorithms regarding the used metrics. Finally, in order to illustrate the application of the five algorithms, a fictitious and simplified example of a civil construction project, using steel structures, is proposed. Based on the results obtained by the algorithms for two scenarios of the proposed example, an analysis about the influence of resource availability with respect to adopted objectives is presented.