Florian Niedermann

Business Process Optimization Using Formalized Optimization Patterns

The success of most of today’s businesses is tied to the efficiency and effectiveness of their core processes. Yet, two major challenges often prevent optimal processes: First, the analysis techniques applied during the optimization are inadequate and fail to include all relevant data sources. Second, the success depends on the abilities of the individual analysts to spot the right designs amongst a plethora of choices. Our deep Business Optimization Platform addresses these challenges through specialized data integration, analysis and optimization facilities. In this paper, we focus on how it uses formalized process optimization patterns for detecting and implementing process improvements.

Supporting manufacturing design by analytics, continuous collaborative process improvement enabled by the advanced manufacturing analytics platform

The manufacturing industry is faced with global competition making efficient, effective and continuously improved manufacturing processes a critical success factor. Yet, media discontinuities, the use of isolated analysis methods on local data sets as well as missing means for sharing analysis results cause a collaborative gap in Manufacturing Process Management that prohibits continuous process improvement. To address this challenge, this paper proposes the Advanced Manufacturing Analytics (AdMA) Platform that bridges the gap by integrating operational and process manufacturing data, defining a repository for analysis results and providing indication-based and pattern-based optimization techniques. Both the conceptual architecture underlying the platform as well as its current implementation are presented in this paper.

BIAEditor : matching process and operational data for a business impact analysis

A profound analysis of all relevant business data in the company is necessary for optimizing business processes effectively. Current analyses typically exclusively run on business process execution data or on operational business data stored in a data warehouse. However, to achieve a more informative analysis and to fully optimize a companys business, a consolidation of all major business data sources is indispensable. Recent matching algorithms are insufficient for this task, since they are restricted either to schema or to process matching. Our demonstration presents BIAEditor that allows to annotate and match process variables and operational data models in order to perform such a global business impact analysis.

Deep Business Optimization: Making Business Process Optimization Theory Work in Practice

The success of most of today’s businesses is tied to the efficiency and effectiveness of their core processes. This importance has been recognized in research, leading to a wealth of sophisticated process optimization and analysis techniques. Their use in practice is, however, often limited as both the selection and the application of the appropriate techniques are challenging tasks. Hence, many techniques are not considered causing potentially significant opportunities of improvement not to be implemented. This paper proposes an approach to addressing this challenge using our deep Business Optimization Platform. By integrating a catalogue of formalized optimization techniques with data analysis and integration capabilities, it assists analysts both with the selection and the application of the most fitting optimization techniques for their specific situation. The paper presents both the concepts underlying this platform as well as its prototypical implementation.

Design-Time Process Optimization through Optimization Patterns and Process Model Matching

The goal of process design is the construction of a process model that is a priori optimal w.r.t. the goal(s) of the business owning the process. Process design is therefore a major factor in determining the process performance and ultimately the success of a business. Despite this importance, the designed process is often less than optimal. This is due to two major challenges: First, since the design is an a priori ability, no actual execution data is available to provide the foundations for design decisions. Second, since modeling decision support is typically basic at best, the quality of the design largely depends on the ability of business analysts to make the ”right” design choices. To address these challenges, we present in this paper our deep Business Optimization Platform that enables (semi-) automated process optimization during process design based on actual execution data. Our platform achieves this task by matching new processes to existing processes stored in a repository based on similarity metrics and by using a set of formalized best-practice process optimization patterns.

Warehousing manufacturing data: a holistic process warehouse for advanced manufacturing analytics

Strong competition in the manufacturing industry makes efficient and effective manufacturing processes a critical success factor. However, existing warehousing and analytics approaches in manufacturing are coined by substantial shortcomings, significantly preventing comprehensive process improvement. Especially, they miss a holistic data base integrating operational and process data, e. g., from Manufacturing Execution and Enterprise Resource Planning systems. To address this challenge, we introduce the Manufacturing Warehouse, a concept for a holistic manufacturing-specific process warehouse as central part of the overall Advanced Manufacturing Analytics Platform. We define a manufacturing process meta model and deduce a universal warehouse model. In addition, we develop a procedure for its instantiation and the integration of concrete source data. Finally, we describe a first proof of concept based on a prototypical implementation.

Automated Process Decision Making Based on Integrated Source Data

Decision activities are frequently responsible for a major part of a process’s duration and resource consumption. The automation of these activities hence holds the promise of significant cost and time savings, however, only if the decision quality does not suffer. To achieve this, it is required to consider data from diverse sources that go beyond the process audit log, which is why approaches relying solely on it are likely to yield sub-optimal results. We therefore present in this paper an approach to process decision automation that incorporates data integration techniques, enabling significant improvements in decision quality.

Beyond Roles: Prediction Model-Based Process Resource Management

The outcome of a business process (e.g., duration, cost, success rate) depends significantly on how well the assigned resources perform at their respective tasks. Currently, this assignment is typically based on a static resource query that specifies the minimum requirements (e.g., role) a resource has to meet. This approach has the major downside that any resource whatsoever that meets the requirements can be retrieved, possibly selecting resources that do not perform well on the task. To address this challenge, we present and evaluate in this paper a model-based approach that uses data integration and mining techniques for selecting resources based on their likely performance for the task or sub-process at hand.

Business impact analysis--a framework for a comprehensive analysis and optimization of business processes

The ability to continuously adapt its business processes is a crucial ability for any company in order to survive in today’s dynamic world. In order to accomplish this task, a company needs to profoundly analyze all its business data. This generates the need for data integration and analysis techniques that allow for a comprehensive analysis.A particular challenge when conducting this analysis is the integration of process data generated by workflow engines and operational data that is produced by business applications and stored in data warehouses. Typically, these two types of data are not matched as their acquisition and analysis follows different principles, i.e., a process-oriented view versus a view focusing on business objects.To address this challenge, we introduce a framework that allows to improve business processes considering an integrated view on process data and operational data. We present and evaluate various architectural options for the data warehouse that provides this integrated view based on a specialized federation layer. This integrated view is also reflected in a set of operators that we introduce. We show how these operators ease the definition of analysis queries and how they allow to extract hidden optimization patterns by using data mining techniques.

A provenance-aware service repository for EAI process modeling tools

One of the major challenges for Enterprise Application Integration (EAI) process modeling tools is the continuous adaptation of the business processes and services. Business and IT specialists are both confronted with a number of problems involved in the adaptation of such processes, such as the lack of support for process lifecycle management, data and functional interoperability problems or the appropriate service knowledge base. Currently, most service engineering methods adopt a lifecycle strategy for the design, implementation, deployment and evaluation of services. However, enterprises exploiting service reusability lack the knowledge on process dependencies across the entire service lifecycle. This knowledge is required by process modeling tools in order to keep EAI processes loosely-coupled. Using a provenance data model we describe the different types of service dependencies in EAI processes with regard to the service changes across its lifecycle. We present a provenance-aware service repository with provenance subscription capabilities and its adoption for different use cases in the manufacturing domain.