Home | Business News | Browse by Publication | A | Architectural Science Review

An improved cost optimisation algorithm for site building layout planning.

Article Excerpt
Abstract: It is important to consider the layout of site facility building in order to evaluate costs for engineering projects. Alternative site building layouts can be optimised in the planning stages before a choice on the most practical and efficient design is chosen. Usually, the least cost for different site building layouts will be an important consideration in the planning process. In this paper, a searching strategy of multi-element algorithm was developed to solve the building layout planning problem. This study indicates that the multi-element search algorithm with tabu search technique has a significant impact in the design layout. A comparison is also presented to demonstrate the efficiency of multi-element search algorithm in solving the site building planning. The performance of the proposed algorithm is more efficient than the genetic algorithms (GA) and annealed neural networks (ANN) in the determination of site building layouts.

Keywords: Cost optimisation, Multi-element search, Site building layout, Tabu search

Introduction

One of the major tasks in the preparation of an engineering construction project is the layout of facility buildings on site. Usually this entails an initial survey of the site to prepare plans for the location of site boundaries and buildings such as job offices, false-work shops, workers' accommodation, material storerooms, steel-reinforcement shops. One must also ensure a distance between buildings that is suitable for traffic. Complete plans with an appropriate layout are required before the commencement of construction to prevent unnecessary costs. Site layout planning has a long history of research in construction cost optimisation. Several different planning algorithms have been developed to optimise the layout costs. These include matching based interactive facility layouts Montreuil, Ratliff, and Goetschalckx, (1987) and a variety of recent algorithms that use a pre-cast yard layout arrangement Cheung, Tong, and Tam, (2002). Researchers have experimented with different methods of domain representation to achieve suitable results. Faced with a variety of different planning algorithms, some planning researchers have been increasingly curious to compare different planning methodologies. It has been a common belief that least-cost planning is the most efficient planning strategy for most planning problems. This belief is based on evidence that least-cost planners can efficiently handle planning problems that involve difficult planning (e.g., Bozer, Meller, & Erlebacher, 1994; Meller, 1996; Tam, 1992).

Some other research on site layout algorithms can be classified in positions of objects in relation to others or the view on objects from others (e.g., Montreuil et al., 1987; Zouein & Tommelein, 1999). Severe efficiency problems with premature convergence to the near-optimal solutions have been found. Some have developed methods using artificial intelligence techniques when conventional methods have proved unsatisfactory. For example, the number of 10 facility buildings will reach at 10! possible layout alternatives. In other words, it is difficult to use the conventional methods in solving the placement alternatives above 3,628,000 (=10!) layouts. Yeh has employed Annealed Neural Networks (ANN) to solve the construction site layout problem (Yeh, 1995). Twelve facilities were to be positioned among twelve possible sites. A penalty coefficient was adopted to accelerate the search, though the choice of parameter made this difficult. The temperature cool coefficient and initial probability coefficient variables were then adjusted in order to reduce the value of the objective function in ANN. This, however, only presented a local optimal solution and did not obtain the global optimum layout.

Heuristic techniques are also employed for large and non-linear problem. Genetic algorithms (GA) can be found in solving site building layout problem. Cheung et al., (2002) used the swap method to represent crossover and mutation procedures through the genetic algorithms in search for the least cost for pre-cast yard layout arrangement. Although Cheung's research successfully solved the pre-cast layout problem by the use of complicated GA with computer programming, it was still not the global optimum layout. Hegazy and Elbeltagi (1999) also used the GA technique to develop an evolution-based site layout-planning model. It included the changes in facility positioning area and locations throughout the project life. Hybrid techniques can be found in layout optimisation for some research. Zhang (2002) employed a combination of expert system and neural networks to solve construction site layout problem practically. Other hybrid systems to optimise the location of site layout problem can be found in (Osman, 2003; Zouein, 2002).

Such difficulties with conventional mathematical methods are overcome in this study by the use of multi-element search algorithm with tabu search technique. Two different models in the facility building layout were employed in this research. Experimental examples are used as illustrations. The different results will be compared to those of the GA and ANN.

Site Facility Building Layout Planning

The facility building...

View this article FREE - Now for a Limited Time, try Goliath Business News
Free for 3 Days!