Optimal construction site layout considering safety and environmental aspects

Optimal Construction Site Layout Considering Safety and Environmental Aspects Haytham M. Sanad1; Mohammad A. Ammar2; and Moheeb E. Ibrahim3 Abstract: A good site layout is vital to ensure the safety of the working environment and effective and efficient operations. Site layout planning has significant impacts on productivity, costs, and duration of construction. Construction site layout planning involves identifying, sizing, and positioning temporary and permanent facilities within the boundary of the construction site. Site layout planning can be viewed as a complex optimization problem. Although construction site layout planning is a critical process, systematical analysis of this problem is always difficult because of the existence of a vast number of trades and interrelated planning constraints. The problem has been solved using two distinct approaches: Optimization techniques and heuristics methods. Mathematical optimization procedures have been developed to produce optimal solutions, but they are only applicable for small-size problems. Artificial intelligent techniques have been used practically to handle real-life problems. On the other hand, heuristic methods have been used to produce good but not optimal solutions for large problems. In this paper, an optimization model has been developed for solving the site layout planning problem considering safety and environmental issues and actual distance between facilities. Genetic algorithms are used as an optimization bed for the developed model. In order to validate the performance of the developed model, a real-life construction project was tested. The obtained results proved that satisfactory solutions were obtained. DOI: 10.1061/共ASCE兲0733-9364共2008兲134:7共536兲 CE Database subject headings: Construction sites; Occupational safety; Environmental issues; Optimization models; Computation. Introduction In the construction industry, site planning is the most overlooked aspect by site engineers. The attitude of the engineers has been that site layout planning will be done as the project progresses. It is generally acknowledged that an efficient overall layout plays a key role in the operational efficiency, cost, and quality of construction. Site layout is routinely performed by managers on construction sites. Yet, project managers usually learn to accomplish this task by trial and error in the course of years of fieldwork. The site layout planning problem is generally defined as the problem of identifying the number and size of temporary facilities 共TFs兲 to be laid out, identifying constraints between facilities, and determining the relative positions of these facilities that satisfy constraints between and allow them to function efficiently 共Zouein et al. 2002兲. TFs are those facilities that serve the construction site but are not being considered a physical part of the structure that is required to be built. Examples of TFs are material stores, fabrication yards, lay-down areas, parking lots, offices, and warehouses. In current practice, site layout planning is often done by adjusting previous plans based mainly on the project manager’s experience and common sense 共Cheng and O’Connor 1996兲. Most construction sites involve some or all of the following items: Basic structures to be constructed, facilities that serve the whole project, obstacles such as trees or old buildings, movement 共internal兲 routes, and site boundaries, even there are no definite barriers. Numerous techniques have been proposed to uncover solutions for site layout problems, but it is very difficult to obtain the optimal one by hand calculations. Therefore, optimization techniques are usually used to search for solutions for site layout problems. In almost all optimization approaches for site layout planning, the layout goal to be attained is to minimize 关interfacilities transportation cost兴. One of the common formulas used to achieve this site layout planning goal is n−1 1 Assistant Lecturer, Structural Engineering Dept., Faculty of Engineering, Tanta Univ., Tanta 31521, Egypt. E-mail: sanadhaytham@ yahoo.com 2 Associate Professor, Structural Engineering Dept., Faculty of Engineering, Tanta Univ., Tanta 31521, Egypt 共corresponding author兲. E-mail: mamammar@yahoo.com 3 Professor, Structural Engineering Dept., Faculty of Engineering, Cairo Univ., Cairo, Egypt. Note. Discussion open until December 1, 2008. Separate discussions must be submitted for individual papers. To extend the closing date by one month, a written request must be filed with the ASCE Managing Editor. The manuscript for this paper was submitted for review and possible publication on October 25, 2006; approved on January 22, 2008. This paper is part of the Journal of Construction Engineering and Management, Vol. 134, No. 7, July 1, 2008. ©ASCE, ISSN 0733-9364/2008/ 7-536–544/$25.00. Min n 兺 dijRij 兺 i=1 j=i+1 共1兲 in which n⫽total number of TFs; dij⫽travel distance between facilities i and j; and Rij⫽parameter that represents: 共1兲 transportation cost between facilities i and j or 共2兲 a relative proximity weight that reflects the required closeness between facilities i and j. Literature Review Many research efforts have been made toward developing a realistic model for the site layout planning problem. The problem has been solved by researchers using two distinct techniques: Math- 536 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2008 ematical optimization and heuristic methods. Mathematical optimization procedures have been designed to produce optimum solutions. Heuristic methods, on the other hand, used to produce good but not optimal solutions. Heuristic methods are based on knowledge-based expert system, and recently on artificial intelligent concepts. However, the first category cannot be adopted for large-scale projects because of the need for huge calculations and efforts. Therefore, the second category is the only practical available means for handling complex real-life projects. Cheng and O’Connor 共1996兲 developed an automated site layout system called ArcSite, which is comprised of a geographical information system integrated with a database management system. A knowledge-based expert system is used to model the spatial requirements of a TF on the construction site. A proximity index is used to determine the optimal site layout, which is calculated based on the number of trips and the attract/repel relationships between two facilities. As a notable limitation to this model, ArcSite can locate only a limited number of facilities. Zouein and Tommelein 共1999兲 developed a model for dynamic layout planning of construction facilities. They used a hybrid incremental solution method, which creates a sequence of layouts that span the entire project duration. The model is solved using linear programming to find out the optimal position so as to minimize transportation and relocation costs. The primary advantage of this method is the possibility of formulating and solving the layout problem using linear programming without undue computations. In addition, the quality of the layout based on traveling cost and relocation cost can be assessed. However, construction facilities can be represented as rectangles only. Also, they have fixed dimensions and can be positioned at 0° 共or 90°兲 orientations only. Li and Love 共1998兲 introduced a genetic algorithm-based 共GA-based兲 model to solve the facility layout problem. The objective of the model is to minimize the total traveling distance between facilities. The problem is described as allocating a set of predetermined facilities into a set of predetermined places, simultaneously satisfying layout constraints and requirements. However, it is assumed that the size of each predetermined location equals the area of the largest facility. Also, the predetermined locations are represented only as rectangles. Hegazy and Elbeltagi 共1999兲 developed a GA-based model for dynamic site layout planning. The model deals with any irregular user-defined site using a two-dimensional grid. Each facility is modeled as a number of grid units. The area of a single grid unit is calculated in advance for each site. The main advantages of this model are its applicability for any user-defined site shape and is not limited to rectangular or square facilities. However, the model does not consider safety considerations and environmental aspects. Also, it considers Euclidean 共center-to-center兲 distances between facilities as a method for distance measurement. Osman et al. 共2003兲 developed an automated computer system called EDSLP 共Evolutionary Dynamic Site Layout Planner兲. The system was used for optimally assigning TFs, simultaneously taking into consideration the dynamic nature of construction projects. EDSLP consists of a data input mean, a computer-aided design user interface, and an evolutionary optimization GA engine. The objective function is to minimize transportation and relocation costs. A sequence of layouts spanning the entire project duration is provided. However, the facilities are limited to rectangular shape only. Also, safety and environmental aspects are not considered. Elbeltagi et al. 共2004兲 introduced a model that considers both safety and productivity issues. Also, parts of the constructed space are utilized as TFs to relieve congestion on restricted sites. When a safety concern rises between two facilities, a large negative value is assigned to the closeness weight. As such, the further the distance between them, the lower the layout score 共thus improving the layout兲. However, this method for dealing with safety considerations does not suit all requirements of projects managers regarding safety issues, as will be discussed later. Moreover, the Euclidean distance method is used in measuring distances between facilities. In almost all models used for the site layout planning problem, safety considerations and environmental aspects have been ignored or at least not modeled in an appropriate manner. Also, distances between facilities are not measured properly. These two main limitations lead to impractical modeling of the site layout problem, and hence need modifications to suit real-life situations. In this paper, an optimization model has been developed for solving the site layout planning problem considering safety and environmental issues. A new method for measuring distance between facilities is used, which is named “actual route distance.” The developed model can also support the dynamic nature of construction projects. Because GAs are adopted here as an optimization technique, basic features of GAs will be briefly reviewed first. Optimization by Genetic Algorithms GAs are search algorithms that are based on the natural selection and genetics to search through decision space for optimum solutions. GAs employ a random yet directed search for locating the globally optimal solution. In addition, GAs perform an intelligent search for a solution from a nearly infinite number of possible solutions. Typically, GAs require a representation scheme to encode feasible solutions for optimization problems. Usually, a solution is represented as a linear string called chromosome whose length varies from one application to another. Some measures of fitness 共objective function兲 are applied to construct better solutions. Once the chromosome structure and the objective function are set, the GA evolutionary procedure takes place on a population of parent chromosomes. Three genetic operations are required: Reproduction, crossover, and mutation. Reproduction is the process by which chromosomes with better fitness values receive correspondingly better copies in the new generation. As the total number of chromosomes in each generation is kept constant, chromosomes with lower fitness values are eliminated. The second operator; crossover, is the process in which chromosomes are able to mix and match their desirable qualities in a random fashion. Crossover 共marriage兲 is conducted by selecting two parent chromosomes, exchanging their information, and producing offsprings. The exchange of information between parent chromosomes is done through a random process. Fig. 1 shows a case of double-point crossover, but single-point crossover may also be used. As an opposite to crossover, mutation is a rare process that resembles the process of a sudden generation of an odd offspring that turns out to be genius 共Goldberg 1989兲. The benefit of the mutation process is that it can break any stagnation in the evolutionary process, avoiding local optima. Recently, research shows that GAs are robust and have the capability to efficiently search complex solution space. The robustness of GAs is due to their capabilities to locate the global optimal. Therefore, GAs are less likely to restrict the search to a local optimum compared with point to point movement, or gradi- JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2008 / 537 Fig. 1. Crossover operation to generate offsprings ent descent optimization technique 共Forrest 1993兲. However, the major drawback of GA-based applications is that they require much greater computational time than traditional methods. Safety Considerations and Environmental Aspects Safety is far more than craftsmen wearing hard hats on construction sites. It is a philosophy that identifies and eliminates job site hazards throughout the life cycle of a project. Accident statistics show that construction is one of the most dangerous industries in the world. However, several easily overlooked factors, such as lack of preplanning, inadequate selection of contractors, and lazy attitudes are significant contributors to these statistics 共Hislop 1999兲. On the other hand, a safe tidy site with good safety regulations is likely to be a site with high moral, few disputes, less absenteeism and labor turnover, and better team work. Environmental planning is considered a proven tool for reducing the impacts from any environmental risks. Typical examples of environmental hazards that occur frequently are noisy workshops and facilities that emit gases, vapors, fumes, dusts, mists, dangerous radiation, and any other harmfull substances 共OSHA 1987兲. Both safety and environmental hazards have bad effects on human beings, and therefore their harmful effects should be prevented or at least mitigated. Facilities of high harmful effects, which are positioned adjacent to neighbors, present actual or potential health hazards, especially when those neighbors are hospital, schools, or residential buildings. Therefore, a prespecified area adjacent to those neighbors becomes an essential requirement. These areas will be prevented from being allocated to any TF, and will be referred to as “prohibited area.” Another consideration is the proper “safety zones” around construction areas. These zones protect workers and entities from falling objects. The Uniform Building Code UBC 共1985兲 states that: “at least 10 ft clearance from buildings or structures shall be kept clear, driveways between and around open storage yards shall be at least 15 ft wide and free from accumulation of rubbish, and material stored inside building under construction shall not be placed within 6 ft of any hoist way or inside floor opening.” The third issue regarding environmental consideration is the dangerous interaction between harmful facilities and sensitive facilities. As discussed earlier, harmful facilities are noisy workshops and those emit harmful substances. These types of facilities should be kept a far distance away from sensitive facilities, such as engineers’ offices, workers facilities, and any buildings containing humans. Therefore, a “minimum distance” should be assigned between each two facilities having this action. Because of the increase attention of safety and environmental considerations mentioned before, it must be considered in the site layout planning problem. The important issues, which will be considered are: Prohibited area, safety zones around constructed facilities, and minimum distance between facilities. Fig. 2. Site and facilities representation Developed Site Layout Planning Model In developing a site layout planning model, previously discussed safety and environmental aspects are considered. Also, the distances between facilities are modeled more practically. First, representation of site and facilities are outlined, and types of facilities are presented. Safety and environmental aspects and distances between facilities are modeled in an appropriate manner. Closeness relationships between facilities are defined. Finally, coding site layout optimization in GAs format is presented. Site and Facilities Representation Site and facilities are modeled in the present study using a twodimensional grid. Each grid unit is called a cell, the area of which is user-defined. Hence, any irregular shape of the site can be modeled, as shown in Fig. 2. A facility can be represented in the grid as a number of grid units, which can be calculated using Number of facility units = Facility area Cell area 共2兲 For ease of modeling, each cell has a location reference that can be calculated using Location reference = 共Row position − 1兲 Total columns + Column position 共3兲 Location reference is an expression used to define the position of any facility in the site 共Elbeltagi 1999兲. It is formulated using the column and row boundaries of the whole site. A facility location reference is defined by the location reference of the top-most left-hand cell of the facility. For example, the location reference of Facility A in Fig. 2 is 23 关共3 − 1兲9 + 5兴. Also, location references of Facilities B and C are 52 and 2, respectively 共Fig. 2兲. Facilities can be placed in a site in three ways, starting from the location reference of the facility: horizontal, vertical, and rectangular. In horizontal and vertical arrangement, the user should specify the width of the facility in a number of grid units. Facility A, in Fig. 2, has an area of five cells, width of two cells, and arranged horizontally. Therefore, Facility A occupies four cells 共2 ⫻ 2兲 starting from Location Reference 23 and the remaining cell 共5-4兲 is arranged from the top-most right-hand side. In Fig. 2, Facility B has an area of five cells and width of two cells, but arranged vertically. In this case, Facility B occupies four cells starting from Location Reference 52 and the remaining cell is arranged from the bottom-most left-hand side. In the rectangu- 538 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2008 dij ⱖ Dmin ij 共4兲 where dij⫽Euclidean 共geometric兲 distance between Facilities i and j and Dmin ij⫽minimum euclidean distance between facilities i and j. The minimum distance is user-defined and should be carefully chosen to prevent dangerous interaction between facilities. Fig. 3. Prohibited areas and shortest route distance lar arrangement, the width of the facility is calculated as the integer of the square root of the facility area expressed in a number of cell units. Facility C in Fig. 2 has four grid units and is arranged as a rectangle starting from Location Reference 2. Types of Facilities In this study, the classification of facilities proposed by Elbeltagi 共1999兲 will be adopted. Facilities are classified as temporary facilities and obstacles. Temporary facilities are those facilities that serve the site during construction. TFs may be fixed in place or nonfixed. Fixed facilities are positioned in a fixed location, such as gates, security rooms, and parking lots. On the other hand, nonfixed facilities need to be optimally located in any empty space on the site. Both fixed and nonfixed facilities are maintaining defined closeness relationships with each other. Also, structures or existing buildings will be referred to as permanent entity. Permanent entities are treated as fixed facilities because they maintain defined closeness relationships with other facilities. Obstacles are entities that are fixed in place and not maintaining any defined closeness relationships with other facilities, such as trees or old buildings. Obstacles cannot be removed from the site, so, spaces used by obstacles in the site cannot be allocated to any other type of facility. Modeling Safety Considerations and Environmental Aspects As discussed before, safety considerations and environmental aspects that will be considered in this model are prohibited area, minimum distance, and safety zones. Modeling of these considerations will be discussed here. Prohibited Areas Fig. 3 shows a typical construction site adjacent to a hospital in the bottom right-hand side. In this case, it is important to prevent some facilities that have harmful effects 共such as noise, air pollution, etc.兲 from being positioned adjacent or near to such sensitive entities. Therefore, a prespecified area in the job site must be prevented from being allocated to facilities with harmful effects. In Fig. 3 a prohibited area is represented by dashed cells. Minimum Distance If hazardous facilities have bad effects on users of other TFs, then a minimum distance between such facilities should be prespecified. The distance between such facilities should satisfy the following relationship: Safety Zones Safety zones represent an additional area added to the physical area of the facilities to protect any person who might be injured by the fall of materials, tools, or equipment being raised or lowered. Safety zones should be specified in adequacy with respect to relevant regulations, such as OSHA 共1987兲 and UBC 共1985兲. For example, if the actual area of the facility⫽50 m2 and the appropriate safety zone⫽20 m2, then the facility area⫽50+ 20= 70 m2. Actual Route Distance between Facilities As discussed in the literature, existing site layout models consider a Euclidean distance between facilities. A Euclidean distance between Facilities 1 and 3, in Fig. 3, is represented by the dashed line, which does not represent the actual traveling distance. To specify the actual route distance between facilities, a network of continuous internal routes in the site should be specified first. These routes are usually used for the movement of equipment and labor between facilities within the site. Internal routes are usually specified by project managers, even if these movement routes are not paved, cemented, covered, or lined. Internal routes are usually continuous to facilitate movement of facilities, and hence measuring traveling distance between them. To specify actual route distances between any two facilities, the nearest cell from the internal routes network for each facility has to be determined first. This cell will be referred to as the representative cell, which acts as a representative for that facility. The shortest route distance between Facilities i and j is calculated by summing up the distance between Facility i and its representative cell, Facility j and its representative cell, and the distance between these two representative cells through the internal routes network. In Fig. 3, the shortest route distance between Facilities 1 and 3 equals the summation of d1, d2, d3, and d4. Facilities Closeness Relationships The interrelationships among facilities are decided by the project manager’s preference and involve some degree of fuzziness and ambiguity. Interrelationships between different facilities are usually referred to as closeness relationships, which can be represented by closeness 共proximity兲 weights. A high proximity weight between two facilities means that they share a high level of interaction and accordingly, the distance between them should be small. Evaluation of these weights can be done using either quantitative or qualitative methods. The expertise or desirability of the project manager leads to a subjective verbal proximity relationship between any two facilities. Several attempts have been made to convert these verbal representations to numeric proximity weights between facilities. One of the popular proximity weights used in construction is given in Table 1, as suggested by Elbeltagi 共1999兲. These proximity numbers are determined using fuzzy set theory. The proximity weights given in Table 1 express an exponential relationship with desired closeness. So, if two facilities are required to be JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2008 / 539 Table 1. Closeness Relationship Values Proximity weight Desired relationship 1 6 36 216 1,296 7,776 Undesirable Unimportant Ordinary Important Especially important Absolutely important close to each other, a high proximity weight is specified to force them to be close to each other in the optimization process, and vice versa. Then, the total weighted travel distances 共objective function兲 of a site layout can be determined. To evaluate the goodness of a possible layout 共chromosome兲, Eq. 共1兲 is used as a fitness function, in which dij represents the actual distance between Facilities i and j and Rij⫽proximity weight between them. Once the chromosome structure and objective function are set, the genetic algorithm evolutionary procedure takes place on a population of parent chromosomes. The simplest way to generate that population is randomly. Once the population is generated, the fitness of each chromosome in this population is evaluated using Eq. 共1兲. The developed site layout planning model was implemented on a spreadsheet 共Microsoft 2002兲 because of its simplicity in use and programmability features 共Green et al. 2000兲. A computer program was coded using the macrolanguage of Microsoft Excel to facilitate model application 共Sanad 2006兲. A full pseudocode for the model procedure is shown in the Appendix. Site Layout Optimization Using Genetic Algorithms An optimization-search procedure has been developed using GAs to optimally locate nonfixed TFs on site. The procedure searches for the optimum location of each facility so that closeness relationships and actual distances between facilities are optimally maintained, simultaneously satisfying safety and environmental considerations. Implementing the GAs technique for the problem at hand involves four primary steps: 共1兲 setting the chromosome structure; 共2兲 deciding the chromosome evaluation criterion 共objective function兲; 共3兲 generating an initial population of chromosomes; and 共4兲 selecting an offspring generation mechanism. Chromosome structure has been set as a string of elements; each corresponds to the location of a nonfixed facility, as shown in Fig. 4. In this case, the chromosome length equals the number of nonfixed facilities 共NF兲, which will be arranged within the site boundary. This formation is adopted in the present study. Evaluating the total travel distance of a given site layout involves determining the centroids of all facilities, their representative cells, and then calculating the actual distances between them. Fig. 4. Chromosome formation Case Study The case study selected for the application of the developed model and automated system is “Tanta University Educational Hospital” located in Tanta City, Egypt. The project involves the construction of three multistory buildings, with perimeter fences and three gates as shown in Fig. 5. The site area is 28,500 m2 with an irregular shape. The construction plan of the project requires six permanent facilities 共three buildings and three guard houses兲 and eighteen TFs. These facilities and related data are given in Table 2. Note that the permanent facilities have blank entries in the last two columns. No site layout plan was initially made to compare with. The contractor’s staff depended mainly on their experience in organizing the facilities in site. Therefore, a disorganized site and material handling problems are expected during the progress of the project. Also, the contractor set aside material storage areas around each building to satisfy their individual needs. This, however, resulted in excessive material waste, extra material handling cost, and less maneuverability within the site. The site location is critical because all adjacent buildings are sensitive to the harmful effects, such as noise and dust. These include Tanta University buildings, emergency hospitals, Al- Fig. 5. General site plan for the case study project 540 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2008 Table 2. Facilities Data for Case Study Project Facility number 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Name Type Building 1 Building 2 Building 3 Batch plant 共*兲 Laydown area 1 Cement warehouse Laydown area 2 Labor rest area Offices Scaffold storage yard Carpentry shop 共*兲 Parking Rebar fabrication yard Warehouses Toilets Welding shop 共*兲 Site offices First aid Machine room 共*兲 Tank Guard house 1 Guard house 2 Guard house 3 Laboratory Size Width 共m2兲 共cells兲 Arrangement Fixed 6,575 Fixed 2,950 Fixed 2,250 Nonfixed 600 Nonfixed 600 Nonfixed 400 Nonfixed 400 Nonfixed 300 Nonfixed 300 Nonfixed 300 Nonfixed 200 Nonfixed 200 Nonfixed 200 Nonfixed 200 Nonfixed 120 Nonfixed 120 Nonfixed 80 Nonfixed 40 Nonfixed 40 Nonfixed 40 Fixed 40 Fixed 40 Fixed 40 Nonfixed 35 — — — 4 4 2 2 3 2 3 2 2 3 2 2 1 1 1 1 1 — — — 1 — — — Vertical Horizontal Horizontal Horizontal Horizontal Vertical Vertical Vertical Vertical Horizontal Horizontal Horizontal Horizontal Vertical Vertical Vertical Vertical — — — Vertical Table 3. Minimum Distances between Facilities 共m兲 Facility number 4 8 9 11 16 8 9 11 16 17 18 21 22 23 60 80 — — 80 80 80 80 80 — — 60 60 — — — — — — — 40 40 — — — — — — — — — 60 60 80 80 80 — — — — 60 60 80 80 80 moalmeen club, and some residential buildings located beside the main street. Batch plant, carpentry shop, welding shop, and machine room are considered dangerous facilities from the safety and environment point of view and are designated by an asterisk beside the facility name in Table 2. Therefore, these facilities are prevented from being located in the prohibited area. The minimum distances between facilities are given in Table 3. For example, the distance between batch plant and guard houses should not be less than 80 m. Also, distances between offices and welding shop should not be less than 40 m. The suggested closeness weights between facilities are given in Table 4. It is recommended to use the values given in Table 1. For example, batch plant is absolutely important to be very close to Buildings 1, 2, and 3. The site was represented by spreadsheet cells, the area of which is chosen to be 25 m2. Using symbols in the cells, the site can be easily drawn. Fig. 6 shows an Excel representation of the site elements given in Fig. 5. Also, the prohibited areas adjacent to the neighbor sensitive buildings are shown. The safety zone for each building is specified around its boundary, and fixed facilities are also located. Having entered all the data, the optimization process can start after specifying the required scheduling data. Three milestone dates can be specified to reflect the dynamic nature of the site layout planning problem. The first period starts at the project start date and finishes August 1, 2000. Because the scheduling start date of Building 3 共Facility No. 4兲 is after August 1, 2000, it is not considered in the site layout optimization for that period. The population size and number of offspring chromosomes used are 400 and 200, respectively. The optimum solution for the second period is shown in Fig. 7. The obtained results and layouts were discussed with the field engineer of the project. The discussion showed that the facilities were arranged in appropriate locations which satisfy the closeness relationships and the minimum distances between facilities. Also, these layouts maintain the safety and environmental aspects, especially with the sensitive neighbors to hazardous effects. The case study on hand has been solved before by Elbeltagi et al. 共2004兲 using the procedure discussed in the literature. The corresponding optimal site layout is shown in Fig. 8. Analysis of Fig. 8 shows that some drawbacks can be observed: • Welding shop 共Facility 16兲, which is safely dangerous, is positioned adjacent to Tanta University buildings. In the devel- Table 4. Suggested Closeness Weights 共Rij兲 for the Case Study Project Facility number 4 6 9 12 13 15 16 17 18 19 20 21 24 1 2 3 4 8 9 10 11 12 13 15 16 17 7,776 1 36 1 1,296 1 36 216 36 1 1 1 1 7,776 1 36 1 1,296 1 36 216 36 1 1 1 1 7,776 1 36 1 1,296 1 36 216 36 1 1 1 1 — 7,776 1 1 1 1 1 1 1 1 216 1 216 1 1 1 1 1 1 1 1 36 1 36 1 1 1 1 — 216 1 1,296 1 36 36 1 36 1 216 1 1 1 1 1 1 36 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 36 1 1 1 1 1 1 — 1 1 1 1 1 1 1 36 1 1 1 1 1 — 1 216 1 1 1 1 1 1 1 1 1 1 1 — 1 36 1 1 36 1 1 1 1 1 1 1 1 — 1 1 36 1 1 1 1 1 1 1 1 1 1 — 36 1 36 1 36 JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2008 / 541 Fig. 6. Site map spreadsheet for the case study Fig. 7. Site layout for second phase of case study oped model, this facility is prevented from being allocated in such location using the prohibited area concept. • Machine room 共Facility 19兲, which is safely dangerous, and offices 共Facility 9兲, which are considered sensitive, are positioned adjacent to each other. In the developed model, they are positioned far from other using the minimum distance concept. • No route is available to access parking 共Facility 12兲, except if the safety zone between Building 1 and Facility 10 is used, and this is considered in violation of the main function of the safety zones. Also Facilities 8, 15, 17, 18, and 24 are located adjacent to each other and near Guard Gate 3 共Facility 23兲. This arrangement prevents site access from Gate 3. These two problems are solved in the developed model using the internal route concept. Other case studies were tested using the developed model. It is observed that the well defined closeness weights for TFs among each others and with the permanent facilities greatly affect facilities positions. In addition, the integration of the schedule and layout information makes it easy for the project manager to track and update the layout of the site. Fig. 9 shows the effect of the population size and number of generations on the convergence of the optimum solution. It is observed that the developed system performed poorly with small population sizes even when a big number of generations is used. A number of generations of 100 can be considered sufficient, as shown in Fig. 9. On the other hand, a big population size 共200 or more兲 tends to ensure optimal solutions. Fig. 8. Site layout in first phase for the case study 共Elbeltagi et al. 2004, ASCE兲 542 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2008 Begin: If facilities i and j are fixed Next i Else If Enter minimum distance between facilities i and j 共Dmin ij兲 End if Next i Fig. 9. Effect of population size and number of generations Summary and Conclusions In this paper, a developed optimization model for solving the site layout planning problem was presented. In developing the new model, two main features were introduced. Measuring distance between facilities in the site using actual route distance and incorporating safety considerations and environmental aspects are considered. The developed model was implemented on a spreadsheet 共Excel兲 because of its simplicity in use and programmability features. The program was coded using the macrolanguage of Microsoft Excel, tested, and then experimented. Also, the dynamic nature of the site layout planning problem is considered by integrating the model with MS Project 共Microsoft Corporation 2000兲. In order to validate the performance of the developed model, a real-life construction project with 28,500 m2 site area was tested. The obtained results proved that satisfactory solutions were obtained. Draw site map; Draw obstacles, internal route net, prohibited areas, safety zones and fixed facilities Perform GA optimization; Input population size 共P兲 and number of generations 共G兲 For each chromosome i: calculate fitness 共i兲; For j = 1 to G Randomly select an operation 共crossover or mutation兲; If crossover Select two parents at random Generate offsprings Else If mutation Select one chromosome at random Generate offspring End if Calculate fitness of offspring chromosome If fitness of offspring chromosome is better than the worst chromosome then Replace the worst chromosome by the offspring Next j Check if termination condition is reached End References Appendix: Full Pseudocode for Model Procedure Begin: Enter facilities data; For i = 1 to number of facilities; Enter facility code, name, and area Enter fixed or nonfixed If fixed Next i Else If nonfixed Enter facility width, arrangement type, and prevented or not prevented form being allocated to prohibited areas End if Next i Enter facilities closeness relationships; n⫽number of facilities For i = 1 to n − 1, j = i + 1 to n If facilities i and j are fixed Next i Else If Enter closeness weight 共Rij兲 End if Next i Enter minimum distances between facilities; For i = 1 to n − 1, j = i + 1 to n Cheng, M. Y., and O’Connor, J. T. 共1996兲. “ArcSite: Enhanced GIS for construction site layout.” J. Constr. Eng. Manage., 122共4兲, 329–336. Elbeltagi, E. 共1999兲. “Construction site management.” Ph.D. thesis, Mansoura Univ., Mansoura, Egypt. Elbeltagi, E., Hegazy, T., and Eldosouky, A. 共2004兲. “Dynamic layout of construction temporary facilities considering safety.” J. Constr. Eng. Manage., 130共4兲, 534–541. Green, J., Stephen, B., and Felipe, M. 共2000兲. Excel 2000 VBA programmer’s reference, Wrox, Birmingham, Ala. Forrest, S. 共1993兲. “Genetic algorithms: Principles of natural selection applied to computation.” Science, 261共5123兲, 872–878. Goldberg, D. E. 共1989兲. Genetic algorithm in search, optimization, and machine learning, Addison–Wesley, New York. Hegazy, T., and Elbeltagi, E. 共1999兲. “EvoSite: Evolution-based model for site layout planning.” J. Comput. Civ. Eng., 13共3兲, 198–206. Hislop, R. 共1999兲. Construction site safety: A guide for managing contractors, Lewis, New York. Li, H., and Love, P. 共1998兲. “Site-level facilities layout using genetic algorithms.” J. Comput. Civ. Eng., 12共4兲, 227–231. Microsoft Corporation. 共2000兲. Microsoft Project 2000, Redmond, Wash. Microsoft Corporation. 共2002兲. Excel reference manual. Redmond, Wash. Occupational Safety and Health Administration 共OSHA兲. 共1987兲. Occupational Safety and Health Administrations, construction industry, USDOL, Washington, D.C., 19–26. Osman, H., Georgy, M., and Ibrahim, M. 共2003兲. “An automated system for dynamic construction site layout planning.” 10th Int. Colloquium on Structural and Geotechnical Engineering, Ain Shams Univ., Cairo, Egypt, 1–13. JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2008 / 543 Sanad, H., 共2006兲. “Optimal arrangement of temporary facilities in construction sites.” Ongoing MSc Research, Structural Engineering Dept., Tanta Univ., Tanta, Egypt. Uniform building code 共UBC兲. 共1985兲. International Conf. of Building Officials, 658–660. Zouein, P., Harmanani, H., and Hajar, A. 共2002兲. “Genetic algorithm for solving site layout problem with unequal-size and constrained facilities.” J. Comput. Civ. Eng., 16共2兲, 143–151. Zouein, P., and Tommelein, I. 共1999兲. “Dynamic layout planning using a hybrid incremental solution method.” J. Constr. Eng. Manage., 125共6兲, 400–408. 544 / JOURNAL OF CONSTRUCTION ENGINEERING AND MANAGEMENT © ASCE / JULY 2008