Article Preview
Top1. Introduction
GA is a robust optimization method (Holland, 1975), operating on a population of solutions rather than a single point, thus increases the chance of escaping from a false local optimum point. However, GA can be slow for complex problems and is criticized to be unsuitable for real-time applications. The use of GA in solving demanding complex problems is likely to overwhelm central processing units using software methods. Thus a hardware GA processor is a direct solution to improve the GA computational speeds. Frequent simple operations such as crossover and mutation can be easily implemented on hardware parallel and pipelined architectures (Man, Tang, & Kwong, 2000).
GA has found many useful and practical applications in the field of project management. As advanced technologies evolve, complicated tasks demanding skilled operations and precise execution such as effective resource allocation, automation of parallel processes and estimation of cost, present major difficulties for project managers. Automation as one aspect of building information modeling thus plays a major role in project progressing the construction industry forward into the future. One of the attractive techniques is that GAs have been employed in other fields such as artificial intelligence and signal processing. GA technique was employed to search for an optimal form-making in (Yi, & Malkawi, 2009)’s study. In (Isaac, & Navon, 2009)’s study, GA was used as part of a stochastic model to estimate adaptive cost of several projects.
Major project management issues related to the use of GAs in construction are solving and optimizing cost and time effectiveness. Scheduling repetitive construction projects were performed using a GA to minimize project running cost (Long and Ohsato, 2009). A multi-objective financial decision support model was successfully based on a fuzzy-GA in Chinese state-owned construction firms (Lam, Ning, & Gao, 2009). The GA was effectively used in three modules: (1) roulette wheel selection, (2) stochastic universal sampling with multi-point crossover and single-point mutation, and (3) stochastic universal sampling with adaptive crossover and mutation. A GA was used as an optimizer in an evolutionary fuzzy neural inference model. The optimizer was successfully employed to minimize project running cost using a web-based conceptual technique (Cheng, & Wu, 2009). New support vector machine inference system for construction management using GA was successfully proposed in (Cheng, & Wu, 2009)’s study in which a fast messy GA approach was also employed. GA chromosome, fitness function, and model architecture were designed.
A general overview on how to possibly link construction industry and construction academics to improve automation in the construction field was comprehensively given in (Succar, 2009). Detailed design processes using GA, object-based modeling and other relevant algorithms were also discussed. Project objectives such as cost, time and productivity were shown to be optimized using a GA in employing a multi-objective design model (Gonzalez, Alarcon, & Molenaar, 2009). A hybrid strategy developed using GAs, simulated annealing and quantum simulated annealing techniques for the discrete time-cost trade-off problem (Sonmez, & Bettemir, 2012). A discrete time-cost-environment trade-off problem for large-scale construction systems with multiple modes under fuzzy uncertainty with a multi-objective decision making model was also established (Xu, Zheng, Zeng, Wu, & Shen, 2012). A GA-based method with the consideration in the search process for optimum project duration and/or cost had been proposed (Dong, Ge, Fischer, & Haddad, 2012). A GA-based model can also used to assist with the project selection and auditor assignment process, which can be set up to find the optimal match between project characteristics and auditor expertise from approximately 5.09E+29 possible combinations (Wang, & Kong, 2012). Based on GAs with a matrix approach, a pre-decision variable to suit the precedent relationship of activities in project scheduling problem can avoid modification of traditional GA operations and chromosome structures which will trigger great time consuming for searching solutions (Fung, Huang, & Tam, 2011). In addition, various other information technology applications using GAs were discussed and mentioned (Wong, Xu, Li, Hong, & Shi, 2004; Li, Chen, Yong, & Kong, 2005; Howard, & Byork, 2008; Schlueter and Thesseling, 2009).