MAIIO: Management and Analysis of Intra- and Inter-Organizational Business Processes


Business processes are present in every enterprise. Indeed, since the early 1990s, companies have progressively recognized that their business processes represent a major source of competitive advantage. However, despite the consensus, several issues hamper the exploitation of the advantages that such processes may deliver and their imperative continuous improvement. In this project, we will focus on two of these issues, namely, their optimization, and the provision of support for inter-organizational business relationships. In our opinion, these two aspects are key for the modernization of the industrial sector within the limits of economic and environmental sustainability. The main goal of this project is to contribute to the digital transition of our industrial environment, with a special focus on small and medium enterprises (SMEs) that do not have the tools, the knowledge, or the resources to be part of this reality. The availability of appropriate tools is key to complete such a goal. The main outcome of the current project will be a toolset, integrating existing tools and newly developed ones, providing access to advance technologies for companies at diverse digitalization maturity levels.

Process and Resource Optimization.

Even though the acknowledgement of the relevance of business processes, in most cases, these processes are not explicitly described, which makes their understanding and improvement very difficult. All companies and organizations deal with their business processes, even informally or in an implicit way, but very few manage them properly, and this is a limitation for success. If appropriate software models were available for those processes, one could reason on them to better understand, refactor, and optimize them with respect to different criteria. Indeed, business process optimization is a strategic activity in organizations because of its potential to increase profit margins and reduce operational costs. The first key component of our toolset will include support for process mining and process learning based on state-of-the-art approaches. However, independently of whether our models are automatically generated from log traces or developed by business experts, our first research goal will focus on process and resource optimization.

Optimization of Resource Allocation.

To accelerate convergence, we will guide the search using meta-heuristics, specifically, some variation of hill-climbing and tabu-search. This procedure will allow us not only to look at factors like cost and execution time, but also to others like energy consumption and waste, which are key for today's concern about climate issues. Given the dynamicity and competitiveness of the environments in which companies need to operate, the optimal allocation of resources cannot be seen as something static. The analysis of dynamic strategies for the allocation of resources will also be addressed. The resource allocation must be dynamically updated, continuously, and not only considering their current usage or results, but they must also consider predictions on future whereabouts. We propose developing combined models, based both on the past and on predictions based on traces and expected workloads. We will use classical stochastic processes and long short-term memory neural networks, trained using trace logs, to develop predictive models that can help in the finding of optimal static and dynamic allocation of resources. The information gathered from the above-described different analyses carried out on a process will also allow us to have an overall better understanding of the business process. The obvious next step is to use such information to solve or reduce known problems. We will investigate the identification of smells 一 bottlenecks, synchronization delays, resource overuse, etc. 一 to suggest model refactoring actions. We will rely on our wide experience on model transformations and on experience reports in the literature to develop refactoring actions as transformations that may be automatically executed on the models. We will use a genetic algorithm to suggest and help in the decision of the refactoring.

Inter-organizational Business Processes.

We will focus on the extension of process mining techniques to B2B environments and the extension of business intelligence techniques to process collaborations. Given the increasing importance of value chain collaboration, even more in an industrial environment in which SMEs prevail, business processes need to be more closely aligned across organizational boundaries. However, business process modeling and design must be enhanced and extended to cope with inter-organizational business relationships, including aspects such as clarification of responsibilities, trust, and confidentiality issues. One of the major problems related to inter-organizational communications comes from the lack of a notion of inter-organizational process, and to the use of ad-hoc formats for electronic data exchange. In this project, we will exploit recent advances in B2B standardization for their consideration in inter-organizational business process modelling and workflow management. Based on the OASIS ebXML Collaboration Protocol Profile and Agreement (CPPA), we will develop infrastructure that will allow us to automate the management and analysis of interorganizational business processes. Thus, building upon state-of-the-art process mining techniques, extended for inter-organizational systems, we will be able to conduct business performance analyses through the alignment of business information in electronically exchanged data to business objectives and performance indicators. The global analysis of inter-dependent organizations' processes will allow us to pose the optimization problem as a global one, in which the goals and restrictions of each participant business process may conflict with those of others. We will apply game-theory techniques to provide solutions to the optimization problem by looking at them as collaborating processes that intend to maximize global goals whilst minimizing the resources each one individually uses.

Scalable Composition of Service-based Applications.

Our group is currently working on scalable methods for the composition of service-based applications. Specifically, from a BPEL/BPMN architectural description of an application and service level agreements (SLAs) and requirements (SRLs), we have developed combined methods, mainly based on heuristics, dynamic programming, and genetic algorithms, in which applications of several thousands of services can be optimally deployed in a few seconds. In this project, we plan to exploit some of these results for the dynamic management of processes and inter-organizational processes, and their analysis. Without scalable methods, such tasks can only be used at design time, and with a significant effort. Thus, we are in an optimal situation to also tackle the third of the above issues, and face the automatic management of large applications. With such capabilities we can not only deploy, but continuously monitor and recover from failures or SLA fails, proposing reconfigurations leading to better results.


We expect a significant economic societal and economic impact of our project. At the end of the project, we will have developed new techniques and tools for modelling and reasoning on intra- and inter-organization business processes, which will be freely available and could be used by any European company. Any company may potentially benefit from our results. SMEs where processes are usually not explicitly modelled will be the main target, but any SME, SME conglomerate, or any other company may take advantage of the optimization of processes and resource usage. In all cases, there is a large margin of improvement and cost reduction. During the project, beyond developing the ideas and implementing them as software tools, we will apply our approach to at least a couple of industrial use-cases to show how our approach can be used in practice. We expect that the project will also have relevant scientific and technical outcomes, mainly related to resource analysis techniques, predictions using neural networks, scalable methods for compositionality, and automated management and recovery of service-based applications. Apart from the direct applicability of these and other results of the project on its specific goals, these results may have direct applications on other related fields.

Our Team

The proposal will be implemented by a consolidated team, led by F. Durán and E. Pimentel, that includes well-known researchers, with broad expertise in the different aspects of the proposal. F. Durán has a broad experience in the development of analysis tools, analysis of business processes, model-driven development, and uncertainty. E. Pimentel has a broad experience leading national and international projects in the context of IoT, cloud computing, and smart cities. C. Canal is an expert in service-based development and mobile applications. The research team also includes J. M. Álvarez, who has collaborated with some of us in the past.

Ernesto Pimentel

Principal investigator

Francisco Durán

Principal investigator

Carlos Canal


José María Álvarez