Name: TDiaco (Techniques for diagnosis, dependability and optimization in Business Process Management Systems)
Main researcher: Rafael Martínez Gasca, PhD
Duration: Five years (Jan-2010 – Dec-2014)
Business Process Management Systems (BPMS) are a strategic option in the software development market, that allow organizations to integrate their technologies with people and processes involved in the business development. The automation of Business Process Management (BPM) allow to reduce the related costs and errors, and assures that key processes are executed efficiently and relevant information can be obtained in order to be used for improving them and reduce errors.
With the increasing complexity of the BPMS, it becomes increasingly necessary to provide mechanisms to maintain adequate levels of service. At the same time, an important challenge is the optimization of processes, which is difficult to deal with some success using the traditional methods of design. It is proposed to incorporate fault diagnosis, dependability and process optimization to the typical BPM lifecycle as additional aspects, that have been applied by the research team in other scopes using different techniques, mainly related to Artificial intelligence.
In the first objective, it is proposed the fault diagnosis in business process models and BPMS, and a diagnosability analysis in these systems. We propose that a methodolgy similar to model based diagnosis can be implemented to determine the tasks that are responsible of the error. The problem has been considered before by defining the equivalencies between the business processes and discrete events systems, but without using all the advantages of model-based diagnosis for discrete events systems.
The second objective is referred to the dependability in access control of business process networks (BPN), at the security and high avalaibility levels. Special attention is done to the access control of obligation and the languages that exist for modeling it, because it is one of the less studied aspects in BPM. Another proposed issue is the realization of a P2P infrastructure to facilitate the improvement of business processes.
The third objective of the project will address the multiobjective optimization of business processes, that has been partially studied in BPM, mainly through evolutionary techniques. Here it is considered the use of hybridization techniques, which have already been successfully applied in other areas. It is important to note that a significant similarity has been shown between BPM and the Planning and Scheduling areas of Artificial Intelligence, so it is proposed the adaptation of optimization algorithms of these fields applied to BPM. Lastly, it is considered to obtain robust solutions in BPM, since its increasing in other fields and the importance of achieving greater flexibility in the execution of processes.