Development of a many-objective model based on NSGA-III for construction projects

In recent years, there has been a considerable increase in the number of stakeholders working with different objectives in any construction project. This has necessitated simultaneous achievement of competing objectives, such as: reduction in time, cost, resources, and adverse environmental impact on a project. In order to achieve a balance among these objectives, several multi-objective construction scheduling models have been reported in the literature. However, several challenges and complexities have also been encountered while incorporating and visualizing more than three objectives simultaneously in such models. Some other challenges faced in the model development are: (i) issues related to large non-dominated population, (ii) being computationally expensive, (iii) difficulties in representation of trade-off surfaces, etc. Some of these challenges have been addressed in this work. The aim of this research is to develop a many-objective scheduling model (MOSM). In order to achieve the aim, the following five research objectives have been set out: (i) To develop a qualitative framework for the selection of the most appropriate optimization algorithm for the construction scheduling model. (ii) To develop a many-objective trade-off scheduling model for construction project. (iii)To check the applicability of the developed many-objective scheduling model (MOSM). (iv) To facilitate the activity-wise life cycle costing (LCC) of a construction project in the MOSM model. (v) To develop a many-objective model graphical user interface (MoMGUI) for the MOSM. To achieve the first research objective, an extensive literature survey was carried out for identification of performance parameters and commonly used algorithms in the instant domain. The literature survey led to the finalization of 13 performance parameters and six optimization algorithms. A pairwise comparison using a questionnaire survey involving the six optimization algorithms was carried out and the responses were analyzed using the principles of consistent fuzzy preference relation (CFPR) method. The analysis using CFPR resulted in ranking the six algorithms. The non-dominated sorting genetic algorithm (NSGA) was found to be the most appropriate algorithm for construction scheduling, while integer/linear programming was the least preferred among the six algorithms. Subsequent to the identification of the NSGA as the most apt algorithm, a many-objective scheduling model (MOSM) was developed to cater to the second research objective. The developed model was validated using the two examples available in the literature. To ascertain the applicability of the developed model for the many-objective trade-off, two additional case study examples adopted from the literature were solved to cater to the third research objective. The first case study example, dealt with project objectives of time-cost-environment-resource (TCER), whereas, the second one dealt with time-cost-quality-safety (TCQS). For the first case study example, first the three objective (time-costenvironment [TCE]) trade-off problem was solved using the developed model. Subsequently, the developed model was used to solve by considering an additional objective that of resource moment to make it a four objective problem (time-costenvironment- resource (TCER). This was required as a build-up to demonstrate the implication of adding the fourth objective on the solution obtained for the preceding three-objective problem. The results showed that the developed model was capable of achieving optimal trade-off solutions for the fourth objective, without compromising the other three. Similar results have also been observed in the case when an additional objective that of safety was added into the existing three objective model pertaining to time-cost-quality. This strongly justified the applicability of the developed model in the many-objective trade-off context. The increasing industrialization demands a sustainable model which may benefit the construction industry in general and construction professionals in particular. The research objective 4 included achieving one of the sustainable pillars that is economy. This was attempted by integrating life cycle cost (LCC) in the scheduling model. To integrate the LCC, an activity-wise life cycle cost analysis (LCCA) was performed in a real-life case considering civil, and electrical and mechanical (E&M) items of works. Activity-wise LCC study was undertaken to gain new insight from the stakeholders in the cost-benefit analysis of a particular activity. It also aimed to help stakeholders, concentrate on the previously neglected items of work that had a significant influence on the LCC of a structure. The results showed that the cost share of civil and E&M items of work were 80% and 20% respectively, of the construction cost. However, when LCC was the parameter, these percentages changed to 60% and 40% for civil and E&M items of work. Once the activity-wise LCC was calculated, they were carried forward in the trade-off model. This trade-off model integrated LCC with time. Subsequently, the time-LCC trade-off (TLT) was analyzed with MOSM. Further, the TLT model was compared with the time-cost trade-off model to have an insight into the results. It emerged from the results that TLT has advantages over TCT, to the extent that it offers a more economic method to execute any activity. In the end, a many-objective model graphical user interface (MoMGUI) was developed to address the fifth research objective. The MoMGUI should facilitate easy application of the developed MOSM by practitioners. The model was developed in MATLAB 'app designer'. The MoMGUI was developed with the aim that it should be capable to deal with two objectives, three objectives and four objectives scheduling trade-off problems. The model was checked for its usability under five criteria comprising effectiveness, efficiency, engagement, error tolerance, and ease of learning. It was found to be working satisfactorily. The devised MOSM provides a useful insight for the construction practitioners. This model is capable of preparing an optimal schedule considering two to four construction project objectives, very efficiently. Further, this study offers economic sustainability as an important ingredient which has been incorporated in one of the objectives and integrated into the model. This will encourage the decisionmakers to use this holistic model in the project planning stage. Moreover, it will help stakeholders to ensure a safe working environment, time-bound completion, environmental friendly, economically sustainable, and a quality end product, in any construction project.