Explorations in Parallel Linear Genetic Programming

<p>Linear Genetic Programming (LGP) is a powerful problem-solving technique, but one with several significant weaknesses. LGP programs consist of a linear sequence of instructions, where each instruction may reuse previously computed results. This structure makes LGP programs compact and powerful, however it also introduces the problem of instruction dependencies. The notion of instruction dependencies expresses the concept that certain instructions rely on other instructions. Instruction dependencies are often disrupted during crossover or mutation when one or more instructions undergo modification. This disruption can cause disproportionately large changes in program output resulting in non-viable offspring and poor algorithm performance. Motivated by biological inspiration and the issue of code disruption, we develop a new form of LGP called Parallel LGP (PLGP). PLGP programs consist of n lists of instructions. These lists are executed in parallel, and the resulting vectors are summed to produce the overall program output. PLGP limits the disruptive effects of crossover and mutation, which allows PLGP to significantly outperform regular LGP. We examine the PLGP architecture and determine that large PLGP programs can be slow to converge. To improve the convergence time of large PLGP programs we develop a new form of PLGP called Cooperative Coevolution PLGP (CC PLGP). CC PLGP adapts the concept of cooperative coevolution to the PLGP architecture. CC PLGP optimizes all program components in parallel, allowing CC PLGP to converge significantly faster than conventional PLGP. We examine the CC PLGP architecture and determine that performance</p>

Explorations in Parallel Linear Genetic Programming

<p>Linear Genetic Programming (LGP) is a powerful problem-solving technique, but one with several significant weaknesses. LGP programs consist of a linear sequence of instructions, where each instruction may reuse previously computed results. This structure makes LGP programs compact and powerful, however it also introduces the problem of instruction dependencies. The notion of instruction dependencies expresses the concept that certain instructions rely on other instructions. Instruction dependencies are often disrupted during crossover or mutation when one or more instructions undergo modification. This disruption can cause disproportionately large changes in program output resulting in non-viable offspring and poor algorithm performance. Motivated by biological inspiration and the issue of code disruption, we develop a new form of LGP called Parallel LGP (PLGP). PLGP programs consist of n lists of instructions. These lists are executed in parallel, and the resulting vectors are summed to produce the overall program output. PLGP limits the disruptive effects of crossover and mutation, which allows PLGP to significantly outperform regular LGP. We examine the PLGP architecture and determine that large PLGP programs can be slow to converge. To improve the convergence time of large PLGP programs we develop a new form of PLGP called Cooperative Coevolution PLGP (CC PLGP). CC PLGP adapts the concept of cooperative coevolution to the PLGP architecture. CC PLGP optimizes all program components in parallel, allowing CC PLGP to converge significantly faster than conventional PLGP. We examine the CC PLGP architecture and determine that performance</p>

What defines biomimetic and bioinspired science and engineering?

Biomimicry, biomimesis and bioinspiration define distinctly different approaches for deepening the understanding of how living systems work and employing this knowledge to meet pressing demands in engineering. Biomimicry involves shear imitation of biological structures that most often do not reproduce the functionality that they have while in the living organisms. Biomimesis aims at reproduction of biological structure-function relationships and advances our knowledge of how different components of complex living systems work. Bioinspiration employs this knowledge in abiotic manners that are optimal for targeted applications. This article introduces and reviews these concepts in a global historic perspective. Representative examples from charge-transfer science and solar-energy engineering illustrate the evolution from biomimetic to bioinspired approaches and show their importance. Bioinspired molecular electrets, aiming at exploration of dipole effects on charge transfer, demonstrate the pintail impacts of biological inspiration that reach beyond its high utilitarian values. The abiotic character of bioinspiration opens doors for the emergence of unprecedented properties and phenomena, beyond what nature can offer.

High-definition human visual attention mapping using wavelets

In this work, three novel approaches to detecting visual attention in images are presented. The idea behind detecting areas within images or video that naturally attract a viewer’s attention is based on the concept of generating pre-attentive saliency maps. Saliency, in and of itself, relates to some measure of “conspicuity” in the visual field and is believed to be an important precursor for many tasks in computer vision. One of the proposed methods in this thesis detects salient regions, while the other two detect salient edges. The classical approach to saliency detection proposed by Itti is extended by introducing wavelets as a lossless resizing tool while maintaining the aspect of biological inspiration. In addition to this, the spectral residual method and the frequency tuned method are modified using wavelets to allow for salient edge detection. Tests show that the proposed methods yield results that are not only comparable to leading,cutting-edge methods, but also exceed them in terms of correct and complete object detection as well as noise reduction.

High-definition human visual attention mapping using wavelets

In this work, three novel approaches to detecting visual attention in images are presented. The idea behind detecting areas within images or video that naturally attract a viewer’s attention is based on the concept of generating pre-attentive saliency maps. Saliency, in and of itself, relates to some measure of “conspicuity” in the visual field and is believed to be an important precursor for many tasks in computer vision. One of the proposed methods in this thesis detects salient regions, while the other two detect salient edges. The classical approach to saliency detection proposed by Itti is extended by introducing wavelets as a lossless resizing tool while maintaining the aspect of biological inspiration. In addition to this, the spectral residual method and the frequency tuned method are modified using wavelets to allow for salient edge detection. Tests show that the proposed methods yield results that are not only comparable to leading,cutting-edge methods, but also exceed them in terms of correct and complete object detection as well as noise reduction.

A new mining life for non-feasible mineral deposits?

&lt;p&gt;ROBOMINERS is developing an innovative approach for the exploitation of currently non-feasible mineral deposits. The approach entails the use of a robot-miner - a bio-inspired reconfigurable robot with a modular nature - in a new mining setting where the activities are nearly invisible and where mining presents less socio-environmental constraints, thus contributing to a more safe and sustainable supply of mineral raw materials.&lt;/p&gt;&lt;p&gt;The main aim is to design and develop a robotic prototype that is able to perform mining related tasks in settings including both abandoned, currently flooded mines not accessible anymore for conventional mining techniques; or places that have formerly been explored, but whose exploitation was considered as uneconomic due to the small-size of deposits, or their difficulty to access.&lt;/p&gt;&lt;p&gt;ROBOMINERS&amp;#8217; innovative approach combines the creation of a new mining ecosystem with novel ideas from other sectors, particularly robotics. At this point, work has been done to understand the best methods for the robotminer&amp;#8217;s development in 1) biological inspiration, 2) perception and localisation tools, 3) behaviour, navigation and control, 4) actuation methods, 5) modularity, 6)autonomy and resilience, and 7) the selective mining ability. All these aspects combined aim to provide the robotminer XXI Century tools for mineral exploration and exploitation of (currently) unfeasible deposits.&lt;/p&gt;&lt;p&gt;At the same time, for the vision of a new vision of a mining ecosystem, work is involving studies on 1) developing computer models and simulations, 2) data management and visualisation, 3) rock-mechanical and geotechnical characterisation studies, 4) analysing ground/rock support methods, bulk transportation methods, backfilling types and methods, and 5) sketching relevant upstream and downstream mining industry analogues for the ROBOMINERS concept.&amp;#160;&amp;#160;&lt;/p&gt;&lt;p&gt;After design and development, based on the previously mentioned studies, the robot-miner is set to be tested at targeted areas representatives which include abandoned and/or operating mines, small but high-grade mineral deposits, unexplored/explored non-economic occurrences and ultra depth, not&amp;#160; easily accessible environments. Possible candidates for testing purposes include mines in the regions of Cornwall (UK), mines in the Kupferschiefer Formation (e.g. Poland) or coal mines in Belgium.&lt;/p&gt;&lt;p&gt;When compared to usual mining methods the ROBOMINERS approach shows: 1) no presence of people in the mine, 2) less mining waste produced, 3) less mining infrastructure, 4) less investment, 5) possibility to explore currently uneconomic resources and 6) new underground small-sized mines, practically &amp;#8220;invisible&amp;#8221;. Altogether, ROBOMINERS can contribute to solve some of the main issues that make mining&amp;#8217;s social license to operate so difficult to get in Europe: land-use, environmental limitations, and socio-economic aspects.&lt;/p&gt;