Distributed task scheduling and allocation using genetic algorithms

As complexity and size of projects increase so do the problems associated with the scheduling and management of the design, manufacturing and assembly processes. In the context of large projects the ability to optimise the scheduling and allocation of these processes can also aid in tendering for contract as well as the management of the project itself. Many large projects will be constructed across distributed sites, each with their own capabilities and specific areas of expertise. Multiple sources may be needed to provide skilled personnel, raw materials, specialised components or facilities for the project, even whole sub-systems within a complex project may be contracted out for financial or time reasons. This paper demonstrates how a computational intelligence technique know as the Genetic Algorithm can be used to optimise design, manufacturing and construction schedules for multiple objectives such as minimising cost and time and maximising utilisation. The system generates a number of near-optimal project scenarios from which a single solution can be selected and implemented by the project manager. (C) 1999 Elsevier Science Ltd. All rights reserved.