· · · Institutional Repository

In de huidige praktijk van gebiedsontwikkelingsprojecten vindt het besluitvormingsproces plaats in pluricentrische besluitvormingsarenas, waarin sprake is van interorganisatorische planning. Door de wederzijdse afhankelijkheid tussen de partijen in deze arenas, de daaraan verbonden steeds wisselende partnerschappen en het onbekende eindresultaat, is het sturen van dergelijke projecten een complexe taak. De betrokken partijen (actoren) in deze arenas zijn niet hirchisch ten opzichte van elkaar geordend, waardoor de traditionele, veelal hirchisch geordende plannings- en besluit-vormingsmethodiek uit de bestaande ruimtelijke ordeningsprocessen, niet meer toereikend is. Om het complexe pluricentrische besluitvormingsproces in de huidige gebiedsontwikkelingen te ondersteunen wordt er in dit rapport een nieuw, op het multi-actoren proces gericht, management-instrument geroduceerd. Met dit instrument kunnen de verschillende sub-besluitvormingsprocessen, die in de arenas optreden, in keer als een multi-level netwerk worden gemodelleerd. Dit gebeurt niet centraal maar door de verschillende actoren zelf. Dit multi-level netwerk bestaat daardoor uit de betrokken actoren, hun activiteiten en hun (besluitvormings)relaties. Als resultaat geeft dit instrument elke actor inzicht in het besluitvormingsnetwerk en de mogelijkheid om in zowel zijn/haar eigen organisatie als in zijn/haar persoonlijke werkzaamheden (en bijbehorende activiteiten en relaties) te sturen. __ In urban planning nowadays, the decision-making process takes place in pluricentric decision-making arenas using multi-actor interaction. Due to mutual interdependency between the actors involved and ever-changing partnerships, managing this process is a complex task. Hence, trying to steer and control this process by means of traditional local-government-based hierarchical planning is nearly impossible. To aid this process a special management system for multi-actor urban decision making is introduced in this thesis. The system models the multi-actor process and the decision making activities in this process as a multi-level network, which consists of actors, activities, and the decision-making relations between them. As a result it provides each actor with the possibility to manage and enhance both his organisational and personal networks and the activities which go along with it. Two key system concepts are the dynamic solution space and the equal collaboration structure based on mutual interdependence. In order to implement these ideas, commonly-used critical path algorithms are insufficient, and a Linear Programming model is used instead. An experimental tool in early development is introduced.

At Mechatronics in Medicine (MiM) Laboratory of Imperial College London, a neurosurgical steerable flexible probe ( STING) that is used to access deep brain lesions through curved trajectories is currently being developed. The focus of my research project is mainly on trajectory planning of the flexible probe i.e. investigation on how to increase efficiency and performance of the trajectory planning. Some experiments have been thoroughly done to measure the performance of a well known sampling based path planning method, Reachability-Guided Rapidly-exploring Random Tree (RG-RRT). The first step to improve the performance was to migrate from MATLAB to Python-C++ which yielded 12-13 times performance speedup. Besides taking a close look at the software implementation details, the second step was to improve the algorithm by implementing a waypoint cache and exploiting some parallelization techniques. The parallelization techniques cover multi-core CPU (OR parallel, AND parallel, OR+AND parallel and Manager-Worker) and GPGPU techniques. At the end of my research project, RG-RRT with waypoint cache was experimentally able to reach 4 times performance speedup, while parallelization on multi-core CPU with AND parallel technique has shown the most significant result by obtaining approximately 5 times performance speedup. The other parallelization, which was done through the use of an NVIDIA CUDA-enabled GPU, has successfully obtained 10 times performance speedup. Despite its higher rate of performance speedup, later it was shown that GPGPU technique suffers the most from inefficiency due to I/O bottleneck that is caused by device-host memory transfer.

En-route charging stations allow electric vehicles to greatly extend their range. However, as a full charge takes a considerable amount of time, there may be significant waiting times at peak hours. To address this problem, we propose a novel navigation system, which communicates its intentions (i.e., routing policies) to other drivers. Using these intentions, our system accurately predicts congestion at charging stations and suggests the most efficient route to its user. We achieve this by extending existing time-dependent stochastic routing algorithms to include the battery's state of charge and charging stations. Furthermore, we describe a novel technique for combining historical information with agent intentions to predict the queues at charging stations. Through simulations we show that our system leads to a significant increase in utility compared to existing approaches that do not explicitly model waiting times or use intentions, in some cases reducing waiting times by over 80% and achieving near-optimal overall journey times.