Using component ensembles for modeling autonomic component collaboration in smart farming

Abstract

Smart systems have become key solutions for many application areas including autonomous farming. The trend we can see now in the smart systems is that they shift from single isolated autonomic and self-adaptive components to larger ecosystems of heavily cooperating components. This increases the reliability and often the cost-effectiveness of the system by replacing one big costly device with a number of smaller and cheaper ones. In this paper, we demonstrate the effect of synergistic collaboration among autonomic components in the domain of smart farming-in particular, the use-case we employ in the demonstration stems from the AFar-Cloud EU project. We exploit the concept of autonomic component ensembles to describe situation-dependent collaboration groups (so called ensembles). The paper shows how the autonomic component ensembles can easily capture complex collaboration rules and how they can include both controllable autonomic components (i.e. drones) and non-controllable environment agents (flocks of birds in our case). As part of the demonstration, we provide an open-source implementation that covers both the specification of the autonomic components and ensembles of the use case, and the discrete event simulation and real-time visualization of the use case. We believe this is useful not only to demonstrate the effectiveness of architectures of collaborative autonomic components for dealing with real-life tasks, but also to build further experiments in the domain. CCS CONCEPTS• Computer systems organization → Self-organizing autonomic computing; • Software and its engineering → Domain specific languages; • Applied computing → Agriculture.

Forming Ensembles at Runtime: A Machine Learning Approach

“…such that has to exist if its parent ensemble instance gets instantiated). chargerAssignments.map(_.drone).allDisjoint The more complete version of the example is available in [9] together with details of the ensemble definition and DSL definition. The complete code of the example is available at https://github.com/smartarch/afcens.…”

“…As this state constantly changes, ensembles have to be continuously re-resolved at runtime. This puts hard practical limits on the time needed to complete the ensemble resolution to be in order of seconds (up to minutes), and consequently on the maximum size of the system (which, based on our experiments [9], depends on the complexity of ensembles often limited to a dozen or a few dozens of components). Longer waiting times means that the system cannot flexibly react to ever changing situations in its environment.…”

“…In this paper we report on our initial experiments in this direction. To demonstrate and evaluate the idea, we adopt a use-case inspired by our work in a smart farming project [9]. We use the agent-based simulator of the use-case scenario to draw indicative results pointing to the feasibility of our approach.…”

“…We use the agent-based simulator of the use-case scenario to draw indicative results pointing to the feasibility of our approach. We describe our running example inspired by our work in the smart farming project [9] in Section 2. Then we explain how to recast the problem of ensemble formation to a classification problem in Section 3.…”

Capturing Dynamicity and Uncertainty in Security and Trust via Situational Patterns

“…(now isEqualOrBefore (transport.scheduledArrival plusMinutes 15)) && AccessToMachine, AccessToBrokenMachine, CancellationOfWorkersThatAreLate val factoryTeam = root(new FactorySystem(factory)) 5 6 } Listing 1: Original security specification Below, we overview the parts of the specification important to this paper. Details about the syntax and semantics of DSL for security ensemble specifications are available at [24].…”