Research On Business Process Self-adaptive Framework Based On Context-aware And Multi-agents

The smart devices are popular with the factor about the development of the sensor technology, the decrease of price and the integrity of operation system recent years. More and more institutions and people are enjoying the advantage of smart facilities. Smart devices could computes and reasons the best services to the user with the predefined information and context. However, the increase amount of smart devices, the different data structure from these devices, the different interactive abilities and sensitive abilities make the combination difficult which also restrict the popularity of themselves. Facing the complex environment formed by multi users and multi smart devices, traditional context-aware technologies are inconsiderate in complex relationship between multi users which leads the mismatch about the users and the provided services. Meanwhile, traditional method could only focus to the specific event while smart office has various kinds of environmental events which leads the system cover few application scenes.In order to solve the above problems, a business process self-adaptive framework based on context-aware and multi agents is proposed in this paper. The system could support the complex interaction between multi-users and multi-devices. Also conventional environments and strong customer definition are considered in the design of the system. Firstly, research on the description of environments, organization, task and resource. Secondly, the four kinds of environmental events are studied up on. Finally, the paper proposes these different strategies facing these environmental events in order to make the system more adaptive.The main research contents of this paper are as follows:1) A business process self-adaptive frameworkThe framework can be divided into three layers, including agent layer, business model layer and UI layer. Agents layer contains all the managed objects among the smart offices ranging from person agents to resource agents. Business model layer is responsible for represents the abstract of agents layer. It is also modeling the organization grouped by the user and the tasks they are executing. On this basis, this layer is responsible for scene state model, strategies and reasoning. UI layer is used for the interactive between user and the system.2) Static model of smart office based on meta dataIn order to manager the whole ecosystem among the smart office, the paper divides the elements into person agents and resource agents. Paper also describe the different kinds of organization among persons and tasks based on resources further. We also analyze the reason and description of environmental events based on person agents, resource agents and timeline. Finally, we propose the scene state model based on state space which could easier for the reasoning engine to infer.3) Self-adaptive strategy based on context aware and multi-agentsOn the basis of static meta model, environmental events and scene state model, organization structure oriented strategy, balanced task oriented strategy and resource oriented strategy are proposed facing different events. Combined with custom data from model builder, reasoning engine would help the system more flexible. Finally, the self-adaptive event logs are proposed which could give the business process builder more feedbacks.Finally, according to the method mentioned in this paper, a prototype system of business process based on context aware and multi agents is designed and completed. An example which uses the thesis defense of students is used to verity the prototype system. The preliminary results show that system could support the complex multi agents and business process based on these agents and the research of this paper has some theoretical value and application prospects.