Capacity management of Enterprise Applications (EAs) encompasses critical IT processes that maximize the IT system’s performance while minimizing operational costs. Effective management of EAs capacity can be enhanced through precise anomaly detection and workload forecasting. Given sufficient historical monitoring data of EAs, Machine Learning (ML) algorithms like Isolation Forest and XGBoost can be applied to address anomaly detection and forecasting tasks. However, the performance of these algorithms can strongly depend on the selected hyper-parameter values. Hence, Hyper-Parameter Optimization (HPO) is crucial for successfully adopting ML methods to solve real-world problems. The existing tuning methods like manual tuning, Grid Search, and Random Search often tend to be computationally expensive and time-consuming, making them ineffective for high-dimensional data. On the other hand, recent successes of Deep Reinforcement Learning (DRL) in handling various optimization problems have sparked interest in exploring its potential. Therefore, this work investigates the adaptation of DRL as a novel HPO approach. We present two implementation scenarios deployed in two real-world use cases and compare the results to the state-of-the-art tuning algorithms. The experiments indicate that the DRL algorithms can be adopted as tuning techniques since they demonstrate consistent policy learning and overall reward maximization.

Data Science (DS) has gained increased relevance due to the potential to extract useful insights from data. Quite commonly, this involves the utilization of Machine Learning (ML). The challenging pursuit of developing and productionizing ML models can be supported and automated through MLOps, a specialization of the DevOps paradigm from software development. Therefore, MLOps offers significant potential for DS projects, which are suffering from notable failure rates. Accordingly, this literature review focuses on examining the current state-of-the-art of the publications in this area. Most importantly, the analysis showed that the current MLOps approaches in the literature predominantly emphasize model development and deployment, while organizational aspects (business understanding, evaluation) in a DS project are neglected. As DS project success is not exclusively dependent on technical matters, advancing the MLOps field by bridging the gap between business objectives and the modeling perspective through appropriate frameworks should be pursued in future research.

The article investigates the application of NeuroEvolution of Augmenting Topologies (NEAT) to generate and parameterize artificial neural networks (ANN) on determining allocation and sequencing decisions in a two-stage hybrid flow shop scheduling environment with family setup times. NEAT is a machine-learning and neural architecture search algorithm, which generates both, the structure and the hyper-parameters of an ANN. Our experiments show that NEAT can compete with state-of-the-art approaches in terms of solution quality and outperforms them regarding computational efficiency. The main contributions of this article are: (i) A comparison of five different strategies, evaluated with 14 different experiments, on how ANNs can be applied for solving allocation and sequencing problems in a hybrid flow shop environment, (ii) a comparison of the best identified NEAT strategy with traditional heuristic and metaheuristic approaches concerning solution quality and computational efficiency.

With the current rapid changes in technologies and business demands, the IT service providers have to configure and adapt their IT system landscapes to cope with the dynamic changes. However, the effects of the adopted changes to the Quality of Service (QoS) must be considered in order to sustain the Service Level Agreements (SLA). QoS is strongly characterized by the Non-Functional Requirements (NFRs) of the system. Among these NFRs, performance is considered one of the most dominant factors to be recognized, where performance requirements can have a comprehensive impact on the system. An earlier study has uncovered some of the deficiencies in the current understanding of the employed terminologies, definitions, and classifications regarding performance requirements. Accordingly, it has proposed a clear model that could diminish the misconceptions or the contradictions among the various classifications. In this paper, an interdependency model will be conducted based on the earlier mentioned model Thus, this interdependency model will aim to reveal the complex interdependencies between the declared performance requirements and regulate the interplay between them. Furthermore, to provide a sufficient evaluation of the proposed performance model, an assessment of the model will be presented based on an online survey that was distributed to various IT practitioners and experts from different IT-engaged organizations.

Many organizations try to apply analytics in order to improve their business processes. More and more cloud services are offered to support these efforts. However, the support of prescriptive analytics is weak. While concepts for such an optimization-as-a-service exist, these require much expert knowledge in solution methods. In this paper, a workflow for optimization-as-a-service is proposed that utilizes an optimization knowledge base in which machine learning techniques are applied to automatically select and parametrize suitable solution algorithms. This would allow consumers to use the service without expert knowledge while reducing operational costs for providers.