Search results for "reinforcement"
showing 10 items of 230 documents
Expanding the Active Inference Landscape: More Intrinsic Motivations in the Perception-Action Loop
2018
Active inference is an ambitious theory that treats perception, inference and action selection of autonomous agents under the heading of a single principle. It suggests biologically plausible explanations for many cognitive phenomena, including consciousness. In active inference, action selection is driven by an objective function that evaluates possible future actions with respect to current, inferred beliefs about the world. Active inference at its core is independent from extrinsic rewards, resulting in a high level of robustness across e.g.\ different environments or agent morphologies. In the literature, paradigms that share this independence have been summarised under the notion of in…
Ultimate Strength and Fatigue Durability of Steel Reinforced Rubber Loading Hoses
2010
Loading hoses in an offshore loading buoy system in the North Sea were investigated with respect to extreme load resistance and fatigue durability. Both experimental work and fatigue life analyses were carried out. The FLS test is based on the principle of a service simulation test according to the American Petroleum Institute (API) 17B guidelines. The test results given in number of endured cycles from the laboratory test are scaled to the in-service conditions. Although the life estimate is based on one full scale test only, an attempt has been made to account for the inherent scatter in fatigue life. Furthermore, the results are validated by large test series with small scale test specim…
Numerical analysis of delamination in through-thickness reinforced composite laminates
2009
Composite laminates show a high vulnerability to out-of-plane actions, responsible for localized damage between two adjacent laminae, i.e. delamination phenomenon. A recent technological solution to improve the strength of the composite laminates in the thickness direction consists in inserting through-thickness reinforcement. In this paper, the composite delamination is analyzed in the context of non-linear fracture mechanics by an original two-phase interface model able to describe the anisotropic elastic and post-elastic mechanical response given by the presence of the reinforcement fibres. The two phases (adhesive joint or matrix of the composite and the reinforcement) are characterized…
Designing a multi-layer edge-computing platform for energy-efficient and delay-aware offloading in vehicular networks
2021
Abstract Vehicular networks are expected to support many time-critical services requiring huge amounts of computation resources with very low delay. However, such requirements may not be fully met by vehicle on-board devices due to their limited processing and storage capabilities. The solution provided by 5G is the application of the Multi-Access Edge Computing (MEC) paradigm, which represents a low-latency alternative to remote clouds. Accordingly, we envision a multi-layer job-offloading scheme based on three levels, i.e., the Vehicular Domain, the MEC Domain and Backhaul Network Domain. In such a view, jobs can be offloaded from the Vehicular Domain to the MEC Domain, and even further o…
Estimation of the tensile strength of an oriented flax fiber-reinforced polymer composite
2011
Unidirectional orientation of natural fibers in a polymer composite ensures the highest efficiency of reinforcement. Flax fiber reinforcement is discontinuous due to limited fiber length and heterogeneous due to the presence of elementary fibers and their bundles. In order to assess the upper limit of tensile strength of such slightly misoriented, nominally UD natural fiber composite, a statistical strength model of continuous UD fiber reinforced composites is applied. It is found that the experimental strength of UD flax composites, produced from rovings or manually aligned fibers, approaches the theoretical limit only at relatively low fiber volume fraction ca. 0.2, being markedly below i…
Realizing Undelayed N-step TD prediction with neural networks
2010
There exist various techniques to extend reinforcement learning algorithms, e.g., eligibility traces and planning. In this paper, an approach is proposed, which combines several extension techniques, such as using eligibility-like traces, using approximators as value functions and exploiting the model of the environment. The obtained method, ‘Undelayed n-step TD prediction’ (TD-P), has produced competitive results when put in conditions of not fully observable environment.
Minimum bond length and size effects in FRP–substrate bonded joints
2009
Abstract The load transfer mechanism between the fibre-reinforced polymer (FRP) materials and the substrate plays a crucial role in the overall response of retrofitted structural members. The FRP–support material interface can be studied by using pull tests in which a reinforcement plate is bonded to a prism and subjected to a tensile force. The experimental results obtained regard the assembly (FRP strip-support block), then a central problem is how to carry out the interface constitutive laws and the related parameters. The principal objective of the present paper is to contrive a procedure which, for a fixed interface constitutive law, permits to derive the interface mechanical parameter…
Strength of sexual and postmating prezygotic barriers varies between sympatric populations with different histories and species abundances
2019
The impact of different reproductive barriers on species or population isolation may vary in different stages of speciation depending on evolutionary forces acting within species and through species’ interactions. Genetic incompatibilities between interacting species are expected to reinforce prezygotic barriers in sympatric populations and lead to cascade reinforcement between conspecific populations living within and outside the areas of sympatry. We tested these predictions and studied whether and how the strength and target of reinforcement between Drosophila montana and Drosophila flavomontana vary between sympatric populations with different histories and species abundances. All barri…
AI for Resource Allocation and Resource Allocation for AI: a two-fold paradigm at the network edge
2022
5G-and-beyond and Internet of Things (IoT) technologies are pushing a shift from the classic cloud-centric view of the network to a new edge-centric vision. In such a perspective, the computation, communication and storage resources are moved closer to the user, to the benefit of network responsiveness/latency, and of an improved context-awareness, that is, the ability to tailor the network services to the live user's experience. However, these improvements do not come for free: edge networks are highly constrained, and do not match the resource abundance of their cloud counterparts. In such a perspective, the proper management of the few available resources is of crucial importance to impr…
The Dreaming Variational Autoencoder for Reinforcement Learning Environments
2018
Reinforcement learning has shown great potential in generalizing over raw sensory data using only a single neural network for value optimization. There are several challenges in the current state-of-the-art reinforcement learning algorithms that prevent them from converging towards the global optima. It is likely that the solution to these problems lies in short- and long-term planning, exploration and memory management for reinforcement learning algorithms. Games are often used to benchmark reinforcement learning algorithms as they provide a flexible, reproducible, and easy to control environment. Regardless, few games feature a state-space where results in exploration, memory, and plannin…