Resource-constrained reasoning using a reasoner composition approach

Abstract

To increase the interoperability and accessibility of data in sensor-rich systems, there has been a recent proliferation of the use of Semantic Web technologies in sensor-rich systems. Quite a range of such applications have emerged, such as hazard monitoring and rescue, context-aware computing, environmental monitoring, field studies, internet of things, and so on. These systems often assume a centralized paradigm for data processing, which does not always hold in reality especially when the systems are deployed in a hostile environment. At runtime, the infrastructure of systems deployed in such an environment is also prone to interference or damage, causing part of the infrastructure to have limited network connection or even to be detached from the rest. A solution to such a problem would be to push the intelligence, such as semantic reasoning, down to the device layer. A key enabler for such a solution is to run semantic reasoning on resourceconstrained devices. This paper shows how reasoner composition (i.e. to automatically adjust a reasoning approach to preserve only a "well-suited" amount of reasoning for a given ontology) can achieve resource-efficient semantic reasoning. Two novel reasoner composition algorithms are introduced and implemented. Evaluation indicates that the reasoner composition algorithms greatly reduce the resources required for OWL reasoning, potentially facilitating greater semantic reasoning on sensor devices.

Benchmarking semantic reasoning on mobile platforms: Towards optimization using OWL2 RL

“…Consequently, Tai et al [48] describe a "selective rule loading" algorithm to determine this ruleset. As a type of naïve forward-chaining algorithm, it executes each rule sequentially on the initial dataset, adding any inferences.…”

“…Implementing and benchmarking this as an optimization for all rules is considered future work. Tai et al [48] propose a selective rule loading algorithm, which composes an OWL2 RL ruleset depending on the input ontology. In our benchmarks, we found that this domain-based rule selection can significantly improve performance.…”

“…Regarding resource-constrained systems however, it has been observed that OWL2 DL is too complex and resource-intensive to achieve scalability [11], [26]. Reflecting this, most mobile semantic reasoners i.e., tailored to resourceconstrained systems, instead focus on rule-based OWL axiomatizations, such as custom entailment rulesets [1], [29] or OWL2 RL rulesets [37], [48]. Indeed, OWL2 RL is an OWL2 profile with a stated goal of scalability, partially axiomatizing the OWL2 RDF-based semantics as a set of rule axioms.…”

“…Indeed, OWL2 RL is an OWL2 profile with a stated goal of scalability, partially axiomatizing the OWL2 RDF-based semantics as a set of rule axioms. Further, a rule-based axiomatization allows easily adjusting reasoning complexity to the application scenario [48] or avoiding resource-heavy inferences [8], [44], by applying subsets of rule axioms. In contrast, transformation rules used in tableau-based DL reasoning are often hardcoded, making it hard to deselect them at runtime [48].…”

“…Further, a rule-based axiomatization allows easily adjusting reasoning complexity to the application scenario [48] or avoiding resource-heavy inferences [8], [44], by applying subsets of rule axioms. In contrast, transformation rules used in tableau-based DL reasoning are often hardcoded, making it hard to deselect them at runtime [48]. Also, most classic DL optimizations improve performance at the cost of memory [11], which is limited in mobile devices.…”