DESIGN FOR ASSEMBLY AND SCHEDULE OPTIMIZATION FOR LARGE STRUCTURE CONSTRUCTION

Abstract

The efficient construction or assembly of a large structure is a critical factor in the success of a project because of the cost and time required. The construction of a large structure is long and expensive due to the number of processes and resources required, the complexity of the processes and the dependencies between tasks. Two methods have been identified for increasing construction efficiency: Design for Assembly (DFA) and schedule optimization. To improve assembly efficiency of smaller products, DFA methodologies have been widely developed and applied. The basis of these methods is to reduce the number of parts in the assembly and to provide a quantitative measure of the assembly efficiency. To achieve similar advantages to those DFA has provided for smaller assemblies, design for large assembly guidelines have been developed. The guidelines are the result of identifying relevant paradigms and the considerations which affect assembly efficiency. The paradigms which have been identified are concurrent engineering, simplicity and modularity. The considerations which must be addressed to improve assembly are joining, handling, 'positioning and assembly planning. To aid designers, the paradigms and considerations have been formulated as guidelines and organized to correspond with the design process. Resource constrained project scheduling is an active area of research focused on reducing project duration and cost. Several optimization techniques have been investigated, including heuristics, branch and bound, genetic algorithms and constraint satisfaction. A new algorithm for resource constrained schedule project optimization has been developed which combines genetic algorithms and constraint satisfaction. The algorithm uses a genetic algorithm to determine the priority of the tasks for resource allocation. Constraint satisfaction techniques are used to generate the schedules. The algorithm has been implemented in a computer program and demonstrated by successfully solving two problems taken from literature.