Towards a Development of Automated-Adaptable Business-Process using Context-Aware and Event-Processing Concepts

Abstract

In the capital markets, things happen fast. The markets move fast, positions change fast. Reacting quickly is the key to increase profit and/or managing risk. Enterprises need to adapt their processes to any observed changes or events in order to maintain their competitive levels and to provide better services. Business processes are static by nature and, therefore, they cannot be modified at run time, and leaving them less useful in the new context. To change the behaviour of a business process, they need to be stopped first, modify it and redeploy it. By doing so, this tends to lose all data associated with the current executions of a running process. To address these issues, this thesis presents a new reference model called Automated-Adapted Business-Process Using Context-Aware and Event-Processing (ABCE) that involves a new approach which allows system developers to design context-aware business processes in which context information can be considered as events monitored in run time. ABCE model developed in this thesis is based on a technique called Complex Event Processing (CEP) which is used to monitor the external working environment and capture relevant events. The captured events are then passed into the system to extract information from them. This extracted information is grouped with other information in order to create a business case with a high-level representation. Depending on the created business case, the adaptation process may starts. Business units are represented as software components which are communicated with each other via key performance indicators (KPIs) interfaces. Each business component (software component) has a set of KPIs with accepted values. The set of KPIs with their accepted values are used to invoke a business component. Business component interfaces are unique. Adaptation process starts on KPI violation. On KPI violation, business process execution changes and a request to appropriate business component will be sent. KPI values which are violated for a business process are accepted for another one. Each business component has one or more business services. When a business component is invoked, one business service is selected according to Service Level Agreement (SLA). After adaptation is completed, the modified version of the business process will be sent for execution to an appropriate execution engine. Finally, we validated our approach by implementing a public transport system using simulation software. The reason behind selecting a public transport system is to increase passengers’ satisfaction by reducing the waiting time for buses and to increase buses owners’ profit. We chose Radio Frequency Identifier (RFID) and Infrared technologies to monitor buses and passengers respectively. By using such technologies the system will be aware about the number of buses and the number of passengers in each passenger track in the station. And by employing supply/demand theory, the system will direct buses to passenger tracks with high demand. Doing so improves the system throughput by up to 40% and reduces passenger waiting time by up to 98%.