30.11.2018

Business process analysis under the microscope of scientists


Nowadays, one of the most popular concepts related to managing a company is the management of business processes. The analysis of business processes has also aroused interest of AGH UST scientists and researchers. As a result, they are going to carry out a research project entitled PACMEL – Process-aware Analytics Support based on Conceptual Models for Event Logs. The project has been granted financing in the international competition dedicated to information and communication technologies “CHIST-ERA Call 2017” within the framework of one of the leading subjects – “Big data and process modelling for smart industry”. The international consortium, whose members are scientific partners from Spain and Italy, is coordinated by the AGH University of Science and Technology. 

The fundamentals of business process management are profound knowledge and the analysis of processes in the context of their continuous improvement. A process analysis cannot take place without the phase of process modelling, i.e. a representation of process events and activities in the form of a model (usually a graphical or mathematical one). Such a model also represents the flow of events, and within the framework of the flow, some possible alternative ways of carrying out the process, parallel activities performed at the same time, etc. 

Within the scope of modelling, analysts use various sources of information, for example, documents, observations and interviews, as well as event logs obtained from computer systems supporting a given process; these event logs are becoming increasingly popular due to the widespread computerisation of company activities and developing interest in advanced data analysis. Such event logs usually contain an ID number of a subsequent process (e.g. customer number, order number), the name of an activity in the process, as well as the time of its completion. 

In the case of multiple companies, obtaining event logs necessary for process modelling is relatively easy thanks to the fact of using different company management systems, for example, ERP, where the results, the times of completion, as well as case and customer numbers are stored. The situation is different, however, in the case of using lower level systems, where a process is not described with the name of a particular activity, but with a number of variables which can have many values, for example, turning a machine on and off, sensors of temperature and movement, etc. In this case, it is necessary to develop conceptual models translating these very simple readings and sensor measurements into event logs containing process phases (activities) that make it possible to develop process modelling at a higher level of generality. In this type of modelling, analysts currently use the techniques of process mining, a dynamically developing discipline of business analytics. 

As a case study, the PACMEL project will use the mining process carried out on a longwall working in a hard coal mine, whose modelling has been worked on for a number of years by a team headed by professor Edyta Brzychczy. On the basis of data obtained from the systems monitoring the operation of the longwall shearer, the scientists will try to develop a general concept of rendering data from sensors into event logs, which can later be used in the modelling of this particular process, as well as other industrial processes. Another aim of the researchers is to develop prototype software that will enable the analysis and visualisation of process models with the use of selected methods and algorithms. 

The project aims at supporting process management in companies by means of developing a possibility of using raw data from an industrial process to its modelling. Information in the form of sensor readings is not often analysed, and even if it is, it is usually analysed in terms of quantity, and with the use of statistical methods or other techniques focusing on data. The work of the team will enable data analysis from a quality- and process-related perspective. 

A particularly important role will be played by the mechanisms of the semantic interpretation of data, its rendering into event logs, and then relating it to a high-level model of the process. It will enable an expert interpretation of the events recorded in a real process and relating them to the concept model.   

Without a process model, it is not only impossible to analyse the process, but there is also no possibility of its re-designing and improvement. On the other hand, such an improvement is difficult without understanding data obtained in a real process, which the project will make possible. The developed methods will be generic in character. It is a reason why the undertaken research is important for the operation of companies in different sectors, especially in the conditions of the so-called “Industry 4.0”. The developed methods will make it possible to draw out knowledge from data stored in IT and computer systems of companies, and use it for the purpose of process improvement. 

At AGH UST, the project is carried out by two research teams: of the Faculty of Electrical Engineering, Automatics, Computer Science and Biomedical Engineering, headed by professor Grzegorz Jacek Nalepa, and of the Faculty of Mining and Geoengineering, headed by professor Edyta Brzychczy. The members of the teams are also professor Marcin Szpyrka and Szymon Bobek, DSc, of the Faculty of Electrical Engineering, Automatics, Computer Science and Biomedical Engineering, and Aneta Napieraj, DSc, of the Faculty of Mining and Geoengineering. The project also offers research grants for doctoral students, who will be chosen through competition. 

PACMEL will be carried out in collaboration with universities in Madrid (Universidad de Autonoma Madrid) and Bozen-Bolzano in Italy (Free University of Bozen-Bolzano). Collaboration with the scientists in Madrid will encompass the methods of intelligent data analysis (data mining, machine learning). The researchers from Italy will deal with the application of semantic methods, including the use of ontology in the modelling of industrial processes. 

The project will last two years. Within the framework of the competition “CHIST-ERA Call 2017”, the project has received the financing of 450 thousand euro.  

The winners of the competition were 14 international research projects (including two projects carried out with the participation of scientific research teams from Poland), with the total budget of more than 12 million euro.