A Business Process Miner for Industry: A Genetic Programming Based Tool

业务流程正变得越来越复杂。管理人员需要对业务流程如何在实际环境中运行有一个准确的了解，并对如何改进流程有指导。一段时间以来，企业资源规划（ERP）软件产品已经能够记录组织的实时托管流程的执行数据。这些数据通常包含有关流程任务及其执行时间的详细信息。可以从显示任务如何链接在一起的数据中手动重建流程流程图；然而，这是一项耗时且容易出错的任务。自动化过程挖掘方法已经由学术团体提出，尽管过程挖掘解决方案的商业实现仍处于非常早期的阶段。虽然越来越多的企业意识到需要自动化流程挖掘技术来进行再工程计划，但当前的实践仍然是由专家驱动的，需要手动检测和解决问题。过程挖掘的目的是从数据日志中识别和提取过程模式，以重构整个过程流程图。数据日志（在业务流程领域通常称为事件日志）包含活动流程的执行数据。这种事件日志可以托管在大中型组织拥有的ERP系统、业务流程管理（BPM）系统和工作流系统中，记录计算机辅助过程的任务完成情况。该建议中概述的技术可以挖掘过程日志并识别过程中的关键特征。也可以挖掘不符合这些特性的流程执行。通过这种方式，可以检测到流程执行方式的变化。这与当前的流程挖掘技术不同，后者的目的是只显示流程的“正确”执行。所建议的技术的功能对于希望对复杂过程流建模并明确识别偏离正常过程执行的组织是有益的。*在线零售商/分析客户在购买商品和服务时必须完成的订购过程。*金融机构/用于通过识别可疑流程执行痕迹来检测欺诈行为。*呼叫中心/检查流程的关键部分是否被忽略。

Business processes are becoming ever more complex. Managers need to have an accurate picture on how a business process is operating in a live environment and guidance on how a process can be improved. For some time Enterprise Resource Planning (ERP) software products have been able to record execution data for an organisation's live hosted processes. Such data typically contains detail on process tasks and the times at which they are executed. It is possible to manually reconstruct a flow chart of a process from this data showing how the tasks link together; however, this is a time consuming and error prone task. Automated process mining methods have been proposed by academic groups though commercial implementation of process mining solutions is still at a very early stage. While there is growing corporate awareness for the need for automated process mining techniques for re-engineering initiatives, current practice is still expert driven, requiring manual problem detection and resolution.The aim of process mining is to identify and extract process patterns from data logs to reconstruct an overall process flowchart. The data logs, more commonly known in the business process field as event logs, contain execution data for a live process. Such event logs may be hosted within ERP systems, Business Process Management (BPM) systems and workflow systems, owned by medium and large organisations, recording the task by task completion of computer assisted processes. The technique outlined in this proposal can mine process logs and identify the key features in a process. Process executions that do not conform to these features can also be mined. In this way, variations in the way a process is executed can be detected. This differs from current process mining techniques that aim to show only the 'correct' execution of a process. The functionality of the proposed technique is of benefit to organisations wishing to model complex process flows and specifically identify departures from normal process execution. Some beneficiaries are:* Online Retailers / to analyse the ordering process a customer must complete in the purchase of goods and services.* Financial Institutions / for the detection of fraud through the identification of suspicious process execution traces.* Call Centres / to check if essential parts of a process are being bypassed.