Research On Offshore Platform Project Scheduling Modeling Technology

Offshore platform project scheduling,which is the most critical part during the platform lifecycle,is a typical cross-organization process that can be fulfilled by the project-oriented and hierarchical supply chain involving the marine enterprise as well as a number enterprises and departments.Whether the tasks can be successfully scheduled directly relates to the construction quality,cost control and time limit quota.Additionally,Multi-Agent System(MAS)technology is a branch of the Distributed Artificial Intelligence(DAI),which shows the distributed,isomeric and collaborative characters that are very suitable to tackle the scheduling problem for offshore platform project.Thus,considering the complexity of offshore platform structures,products and scheduling process,the offshore platform scheduling model was constructed in this paper combined with the characters of cross-parallel scheduling process and agents' collaboration.Firstly,research aspects both at home and abroad about the MAS theory development,task decomposition mechanism,outsourcing scheduling and production scheduling were analyzed and summarized.The MAS technology was adopted to propose the framework of offshore platform project scheduling.Secondly,the decomposition mechanism was studied and the "matrix" decomposition mode was built in terms of pratical characters.Thirdly,based on the Market Mehanism,outsourcing scheduling model was constructed;besides,the decision-making,case-based reasoning and negotiation strategies were studied.Fourthly,aimed at solving the mathematical model for production scheduling,the adaptive hybrid genetic algorithm was designed;moreover,the Bayesian self-learning negotiation strategy was utilized to study the negotiation mechanism between agents so as to realize the allocation of critical resources.Last,by adopting the Java language and B/S structure,the system development was initially realized to verify the modeling feasibility and rationality.