基于Activiti和Drools的水产养殖自动决策流程管理系统

Process management system of automatic decision making for aquaculture based on Activiti and Drools

针对目前水产养殖过程管理中缺乏有效的流程化管理、养殖业务规则不明确、养殖决策效率低、人工经验判断操作失误率高等问题,阐述了工作流技术和规则引擎技术结合应用于水产养殖流程管理的设计思想。首先提出了一种基于规则引擎的智能工作流管理系统体系结构,并对系统进行总体框架设计。然后在分析和总结出水产养殖业务流程步骤繁琐不明确、受时间限制、多循环、多并发和存在大量养殖业务决策等特点的基础上,以南美白对虾养殖流程为例,通过Activiti可视化流程设计器对南美白对虾养殖流程进行流程定义和设计,构建水产养殖工作流模型;通过Drools规则引擎制定南美白对虾养殖关键流程业务规则库,制定过程包括养殖规则的分析、定义、设计、编写和触发。最后,将Activiti工作流引擎和Drools规则引擎与当今主流的SSH框架进行整合,开发基于Activiti和Drools的水产养殖自动决策流程管理系统。通过初期试验运行表明,该系统能有效实现水产养殖全流程业务状态的实时监控和养殖关键流程的规则决策,能有效增加养殖产量14.8%,降低饵料系数11.4%,减少人工管理成本30%,使得整体经济效益增加29.3%,初期试验效果良好,为水产养殖信息化过程管理提供了一种切实可行的流程化、自动化、精细化养殖的新策略。

At present, there exist a series of problems in the process management of aquaculture, including inefficient process management, inexplicit aquaculture business rules, low efficiency of decision-making and high failure rate of operation by human experiences. In order to solve those problems, this study targets to develop an enhanced process management system for the automatic decision making of aquaculture. The theoretical foundation of the proposed system is to integrate rule engine technology with the traditional workflow systems. Although workflow technology can realize the automatic process management and improve the work efficiency, it performs worse in automatic business decision-making. There exist substantial decisions during the process of aquaculture, which can determine the right next-step operation. Rule engine can separate business rules from the application program, and assist workflow system to conduct automatic decision-making and dynamic scheduling of the business process. Based on this theoretical foundation, the architecture of an intelligent workflow management system based on rule engine was proposed, which implemented the idea of interaction and bidirectional driving between workflow engine and rule engine. To further examine the rationale of the proposed system, this study used the case of Penaeus Vannamei breeding to customize the process management system. First, the characteristics of the business process of Penaeus Vannamei breading, including complex and inexplicit steps, being constrained by time, multiple cycles, multiple concurrency and substantial business decisions, were analyzed and summarized. Next, the breeding process of Penaeus vannamei were defined and designed through the Activiti visualization process designer based on the development of workflow model of aquaculture. Then, the analysis, definition, design, compilation and triggering of the business rules in the key breeding processes of Penaeus Vannamei were established through the Drools rule engine. Finally, though integrating Activiti workflow engine and Drools rule engine with the mainstream SSH framework, the process management system of automatic decision making for Penaeus vannamei based on Activiti and Drools was developed by the tool of Eclipse3.7 integrated with the plug-in components of Activiti5.22 and Drools5.6. The process management system followed the three-layer system architecture and programmed by JAVA language and MySQL database. Although compared with the traditional aquaculture management system, the proposed process management system performed superior in process management and automatic decision making, it can still run efficiently. Because the velocity of decision-making by Drools rule engine was not affected by the number of the business rules, no matter how many breeding rules there were, the process management system could always make automatic decisions with an average of 3.5 s and the accuracy rate reached up to 100%. Currently, the process management system had been partially applied by an aquaculture company in Nantong and the company had completed the contrast test in system pilot run. The results of the comparative experiment had demonstrated that the system could effectively realize the real-time supervision and control of the whole-process business state of aquaculture, and quickly and accurately make the automatic decisions of the breeding business rules. Compared with the shrimp pond controlled by the traditional feeding system, the shrimp pond controlled by the process management system increased the yield of aquaculture by 14.8%, reduced the feed coefficient by 11.4%, reduced the labor cost by 30%, and increased the total profit by 29.3%. Thus, this study provides an automatic, streamlined and fine management method for the process management of aquaculture.