90 research outputs found
Robot-learning using a Tree-based Policy Representation
No full textLearning is an important aspect in creating versatile robots. Pre-programming a robot to acquire a wide variety of skills in an ever changing environment is unfeasible. Robot learning provides a promising alternative. Two well-established learning techniques are Programming by Demonstration (PbD) and Learning from Exploration (LfE). PbD and LfE are often combined to strengthen each other. PbD is used because it allows fast learning: with only a few demonstrations, robots are able to reproduce tasks with reasonable performance. After these demonstrations, LfE is used to improve the robot's task performance or to adjust this skills to changing environments. Robots often use continuous mappings between states and actions to represent a skill. Such mappings are called policies and are represented by function approximators. The shape of the policy is determined by the parameters. During learning the robot tries to find optimal parameters for the policy. As the complexity of the skill increases, the number of parameters required to accurately describe the policy for this skill also increases. As the number of parameters increases, the complexity of the solution space increases as well. It is most likely that the LfE algorithm requires more trials to converge for these complex search spaces than simpler search spaces, thus the LfE performance decreases as the complexity of the search space increases. In this thesis a novel multi-resolution policy representation is investigated. The method, called Tree-based policy representation, creates a multi-resolution model based on demonstration data. After this initialization, LfE can use the structure of the Multi-resolution policy to increase learning performance. The method is tested in multiple experimental scenarios. The Tree-based policy representation achieves better learning performance compared to conventional `flat' policy representations, when learning motions that clearly have a multi-resolution aspect. In other cases, the Tree-based movement representation performs equally well or worse compared to standard `flat' policy representations.BMDBioMechanical EngineeringMechanical, Maritime and Materials Engineerin
CO2 reduction as a market opportunity. A product design for Nuon, in order to anticipate the need for CO2 reduction in the business market.
No full textTechnology, Policy and Managemen
Tussen kust en kennis. Analyse van bedreigingen en kansen voor kennisgeneratie met betrekking tot kustveiligheid.
No full textTechnology, Policy and Managemen
The impact of building features on the building's financial performance. "Do pre-crisis and after-crisis features decide?"
No full textThis research studies the link between specific, physical and non-physical, building characteristics and the net rental income generated by the property. Special attention is paid to whether building characteristics that have shown to be decisive in acquisition strategies during the ‘pre-crisis’ period are still a relevant decision-making indicator in the ‘after-crisis’ period.Corporate Real Estate ManagementReal Estate & HousingArchitecture and The Built Environmen
Measuring the Effectiveness of Enterprise Architecture Implementation
No full textTechnology, Policy and Managemen
A possible new adaptive control approach for ramp metering
No full textDue to increasing congestion there is a need for improved traffic control measures. During the Field Operational Test Integrated Management Amsterdam (Praktijkproef Amsterdam (PPA)) several control measures have been tested. One of them is ramp metering. The ramp metering algorithm, ALINEA, has been used with the "Parameterschatter" (Parameter estimator), which tries to estimate the critical density every time step. This is called adaptive control. This critical density is used to update the target density in the ALINEA algorithm. In an attempt to improve the ramp metering algorithm, next to the Parameterschatter, also a gradient method is used. The gradient method tries to tune the parameter gains in the ramp metering algorithm. In this thesis a variation on the normal ALINEA is used, a so-called Proportional-Integral(PI)-ALINEA. This PI-ALINEA has two parameter gains which will be updated towards the current situation on the freeway. The new adaptive controller (AD-RMC) is tested with the macroscopic traffic flow model METANET using a variable critical density and noisy measurements. The conclusion of this thesis is that the tuning of the parameter gains has benefit on the performance of the ramp meter installation in terms of travel time. Further research is needed (with for example a microscopic simulation model) on this algorithm before implementing it in practice.Traffic Flow Theory and Traffic NetworksTransport & PlanningCivil Engineering and Geoscience
Grinded Dutch calcite as seeding material in the pellet softening process
No full textCalcium carbonate pellets are produced as a by-product in the softening process. In The Netherlands, these pellets are applied as a raw material in several industrial and agricultural processes. The sand grain inside the pellet hinders the application in some high potential market segments such as paper and glass. Earlier work showed that it is technologically possible and in some cases economically feasible to replace sand grains with commercially available calcite grains obtained from limestone (e.g. in Italy or Germany). ). In order to further increase sustainability by reduction of transport, it might be possible to grind and sieve the pellets and reuse this grinded Dutch calcite as seeding material for pellet softening process. The goal of this research was to determine the possibilities of the use of grinded Dutch calcite as an alternative for the use of garnet sand as seeding material in the pellet softening process. A comparison between grinded calcite and garnet sand is made based on the five key performance indicators (K.P.I.’s): quality, quantity, costs, reliability and sustainability. Pilot scale tests were conducted at Weesperkarspel (Waternet) during the winter period. The produced calcite pellets were dried, grinded and sieved calcite as seeding material (0.5 mm) by grinding and sieving the Dutch calcite pellets (1 mm). The produced calcite grains were applied as seeding material in a fluidized bed softening reactor (? = 30 cm; height = 6 m; velocity = 60-80 m/h). The effect on sustainability was based on a life cycle analysis (LCA) performed in SimaPro. An FMECA was carried out to identify possible risks failures for the fullscale implementation. The grinding and sieving experiment showed that an overall yield of 40 % of grinded calcite pellets with a diameter of 0.4 – 0.6 mm, could be reached. The pilot pellet softening experiments showed that the softening performance of the Italian calcite and the grinded Dutch calcite scenario, in terms of water quality parameters (total hardness, turbidity, pH, NaOH dosage), were comparable from the garnet scenario. The results of the modeling in Stimela showed comparable results with the pilot plan experiments. The LCA concerning the Weesperkarspel plant and its pre-treatment showed a score of 215,250 EcoPoints (Barrios et al., 2008). The scenario of grinded calcite seeding material results in a decrease of about 3824 EcoPoints, 1.8 %, due to reduced transportation of garnet. Usage of grinded pellets in Dutch industries (compared to imported calcite) results in a total decrease of 10.640 EcoPoints, 5%, mainly caused by the reduced transportation. The business case showed that costs of the seeding material will be lower due to the reuse of the grinded pellets as seeding material. Due to the improved pellet quality the pellet turnover will increase. The worst-case scenario showed a possible cost reduction for the WPK plant of \u80 38,000 (1.8%). For both treatments of Waternet a costs reduction of \u80 120.310 (1%) can be achieved. The scenario study for the full-scale implementation showed that a transition, of all eight reactors at the treatment facility of WPK, from garnet sand to calcite as a seeding material introduced the fewest risks for water quality and reliability. The FMECA showed that for the full-scale implementation the hygiene of the seeding material is an important factor.Sanitary EngineeringWater ManagementCivil Engineering and Geoscience
Verkenningen secundaire grondstoffen: 1996-2015
No full textHet doel van dit project is het verschaffen van kwantitatieve informatie voor de opstellers van het Structuurschema Oppervlatedelfstoffen (SOD II) met betrekking tot het verwachte aanbod aan secundaire grondstoffen in de periode 1996-2015 en de verwachte toepassingsmogelijkheden hiervan. In het rapport wordt een overzicht gegeven van de productie en afzet van secundaire grondstoffen, maar door het gebrek aan gegevens over producten uit bouw- en sloopafval is het trekken van algemene conclusies uit dit cijfermateriaal niet goed mogelijk
Noise minimization on houses around airports
No full textThis thesis was a combined project of the CWI (the national research institute for mathematics and computer science in the Netherlands) and the NLR (National Aerospace Laboratory of the Netherlands). In this thesis we examine noise pollution around airports. Given an airport we will consider the problem maximizing the number of flights, while minimizing the number of houses suffering more than the threshold amount of noise pollution. We will formulate this problem as a multi-objective optimization problem. We will look at the computational complexity of the problem and its approximability. Using the concept of Pareto optimality we will consider methods to rewrite our problem to a single objective optimization problem. Using the methods and the computational complexity and approximability results obtained, we will construct algorithms to solve them problem. Finally we will use these algorithm on problem instances to give insights in their performance.Applied mathematicsElectrical Engineering, Mathematics and Computer Scienc
- …
