Non-Functional Requirements (NFRs) play a critical role in driving self-adaptation in software systems. In Self-Adaptive Systems (SAS), satisfying multiple NFRs simultaneously introduces significant complexity, as these requirements often conflict—improving one NFR can negatively impact others. Addressing such tradeoffs becomes even more challenging due to the varying degrees of observability of NFRs, with some being fully observable and others only partially observable. Traditional approaches to SAS decision-making, such as those based on Markov Decision Processes (MDPs), often assume homogeneous observability, which limits their ability to address these challenges effectively. We argue that treating NFRs as having mixed observability—where some are fully observable and others are partially observable—enables more effective decision-making. How can SAS model and resolve tradeoffs among NFRs with mixed observability to achieve better outcomes? This article introduces SPECTRA, a multi-objective decision framework based on MDPs. SPECTRA addresses tradeoffs among NFRs by leveraging a multi-objective Mixed Observability Markov Decision Process (MOMDP), which models and handles the varying observability of NFRs effectively. The approach is evaluated using scenarios from MirrorNet, a realistic Remote Data Mirroring (RDM) system utilizing Software-Defined Networking (SDN). Results show that SPECTRA achieves higher utility values, faster policy planning, and more effective tradeoffs compared to existing approaches.

Self-Adaptive systems (SAS) must make runtime decisions to balance trade-offs among competing quality-of-service (QoS) requirements — such as cost, performance, and reliability — under uncertain and dynamic conditions. Current approaches to support this, such as Pareto-Based or utility-driven methods, often lack a quantifiable notion of what constitutes an acceptable loss in one requirement in favor of another. Inspired by practices in clinical trials, we propose a novel, requirements-centric application of Non-Inferiority (NI) Trials to decision-making in SAS. We reinterpret the NI margin, traditionally used to determine the acceptability of new treatments, as a stakeholder-specified tolerance threshold for QoS trade-offs. This offers a statistically grounded method to assess whether a new decision-making technique achieves ‘good enough’ performance according to stakeholder-defined thresholds, relative to established alternatives. We apply this approach to compare two reinforcement learning techniques in an SAS context and demonstrate how it captures nuanced trade-off decisions in QoS satisfaction. We argue that NI Trials can complement existing RE methods by introducing a principled mechanism for reasoning about acceptable degradation, supporting negotiation, monitoring, and prioritization of requirements in uncertain environments.

Approaches that support the decision-making of self-adaptive and autonomous systems (SAS) often consider an idealized situation where (i) the system’s state is treated as fully observable by the monitoring infrastructure, and (ii) adaptation actions are assumed to have known, deterministic effects over the system. However, in practice, the system’s state may not be fully observable, and the adaptation actions may produce unexpected effects due to uncertain factors. This article presents a novel probabilistic approach to quantify the uncertainty associated with the effects of adaptation actions on the state of a SAS. Supported by Bayesian inference and POMDPs (Partially-Observable Markov Decision Processes), these effects are translated into the satisfaction levels of the non-functional requirements (NFRs) to, therefore, drive the decision-making. The approach has been applied to two substantial case studies from the networking and Internet of Things (IoT) domains, using two different POMDP solvers. The results show that the approach delivers statistically significant improvements in supporting decision-making for SAS.

Abstract Self-adaptive systems (SASs) are increasingly leveraging autonomy in their decision-making to manage uncertainty in their operating environments. A key problem with SASs is ensuring their requirements remain satisfied as they adapt. The trade-off analysis of the non-functional requirements (NFRs) is key to establish balance among them. Further, when performing the trade-offs it is necessary to know the importance of each NFR to be able to resolve conflicts among them. Such trade-off analyses are often built upon optimisation methods, including decision analysis and utility theory. A problem with these techniques is that they use a single-scalar utility value to represent the overall combined priority for all the NFRs. However, this combined scalar priority value may hide information about the impacts of the environmental contexts on the individual NFRs’ priorities, which may change over time. Hence, there is a need for support for runtime, autonomous reasoning about the separate priority values for each NFR, while using the knowledge acquired based on evidence collected. In this paper, we propose Pri-AwaRE, a self-adaptive architecture that makes use of Multi-Reward Partially Observable Markov Decision Process (MR-POMDP) to perform decision-making for SASs while offering awareness of NFRs’ priorities. MR-POMDP is used as a priority-aware runtime specification model to support runtime reasoning and autonomous tuning of the distinct priority values of NFRs using a vector-valued reward function. We also evaluate the usefulness of our Pri-AwaRE approach by applying it to two substantial example applications from the networking and IoT domains.

The main objective of decision-making in a self-adaptive system (SAS) is to continuously satisfy its requirements under environmental uncertainty. As the run-time context changes, the system may need to re-configure itself by making trade-offs between the non-functional requirements (NFRs) based on their individual priorities for satisfaction. We demonstrate Pri-AwaRE as an approach to support priority-aware decision-making in SASs by providing explicit runtime modelling and reasoning of individual priorities of NFRs. The approach also supports autonomous tuning of the priorities under dynamic situations to maintain the required satisfaction levels of NFRs. In this paper, we showcase how Pri-AwaRE is used in a substantial industrial case of Remote Mirroring using a simulation tool called RDMSim. Our results show that Pri-AwaRE offers the required satisfaction levels of NFRs by autonomously tuning of NFRs’ priorities according to new runtime environmental contexts.

A Self-adaptive system (SAS) is required to continuously satisfy its requirements at runtime under environmental uncertainty. As the run-time context changes, the system may need to re-conFigure itself, and since resources are finite this may require trade-offs between the SAS's non-functional requirements (NFRs). A number of runtime modelling techniques have been developed for resolution of uncertainty in SASs. However, current techniques lack the explicit runtime representation of NFR priorities, leading to the risk that adaptations may fail to respect the NFRs' priorities. In order to address this problem, my research will focus on the explicit representation and a priori elicitation of distinct NFRs' priorities to support autonomous NFR tradeoff decisions. This paper gives an overview of the research area by identifying gaps in existing state of the art techniques. The paper also provides a description of the prospective approaches for solving the problem along with current progress made and next steps.