Minimizing test-point allocation to improve diagnosability in business process models

Abstract

Diagnosability analysis aims to determine whether observations available during the execution of a sys-tem are sufficient to precisely locate the source of a problem. Previous work deals with the diagnosabilityproblem in contexts such as circuits and systems, but no with the adaptation of the diagnosability prob-lem to business processes. In order to improve the diagnosability, a set of test points needs to be allocated.Therefore, the aim of this contribution is to determine a test-point allocation to obtain sufficient observ-able data in the dataflow to allow the discrimination of faults for a later diagnosis process. The allocationof test points depends on the strategies of the companies, for this reason we defined two possibilities: toimprove the diagnosability of a business process for a fixed number of test points and the minimizationof the number of test points for a given level of diagnosability. Both strategies have been implemented inthe Test-Point Allocator tool in order to facilitate the integration of the test points in the business processmodel life cycle. Experimental results indicate that diagnosability of business processes can be improvedby allocating test points in an acceptable time.