Thesis (M. Phil.)--University of Hong Kong, 2010. / Includes bibliographical references (p. 123-130). Also available in print.

In this thesis two case studies are performed about solving two design problems we face during the design phase of new Volvo truck. One is to solve the frame packing problem on CAN bus. The other is to solve the LDC allocation problem. Both solutions are targeted to meet as many end-to-end latency requirements as possible. Now the solution is obtained through manually approach and based on the designer experience. But it is still not satisfactory enough. With the development of artificial intelligence method we propose two methods based on genetic algorithm to solve our design problem we face today. In first case study about frame packing we perform one single genetic algorithm process to find the optimal solution. In second case study about LDC allocation we proposed how to handle two genetic algorithm processes together to reach the optimal solution. In this thesis we show the feasibility of adopting artificial intelligence concept in some activities of the truck design phases like we do in both case studies.

No description available.

No description available.

Access to abstract restricted until 05/2015. / Asscess to thesis restricted until 05/2015. / Department of Computer Science

Thesis (M.S.)--Rochester Institute of Technology, 2009. / Typescript. Includes bibliographical references (leaves 58-59).

This doctoral thesis will present the results of work into optimisation algorithms. We first give a detailed exploration of the problems involved in comparing optimisation algorithms. In particular we provide extensions and refinements to no free lunch results, exploring algorithms with arbitrary stopping conditions, optimisation under restricted metrics, parallel computing and free lunches, and head-to-head minimax behaviour. We also characterise no free lunch results in terms of order statistics. We then ask what really constitutes understanding of an optimisation algorithm. We argue that one central part of understanding an optimiser is knowing its Bayesian prior and cost function. We then pursue a general Bayesian framing of optimisation, and prove that this Bayesian perspective is applicable to all optimisers, and that even seemingly non-Bayesian optimisers can be understood in this way. Specifically we prove that arbitrary optimisation algorithms can be represented as a prior and a cost function. We examine the relationship between the Kolmogorov complexity of the optimiser and the Kolmogorov complexity of it’s corresponding prior. We also extended our results from deterministic optimisers to stochastic optimisers and forgetful optimisers, and we show that uniform randomly selecting a prior is not equivalent to uniform randomly selecting an optimisation behaviour. Lastly we consider what the best way to go about gaining a Bayesian understanding of real optimisation algorithms is. We use the developed Bayesian framework to explore the affects of some common approaches to constructing meta-heuristic optimisation algorithms, such as on-line parameter adaptation. We conclude by exploring an approach to uncovering the probabilistic beliefs of optimisers with a “shattering” method.

Les problèmes d’optimisation combinatoire sont d’un grand intérêt à la fois pour le monde scientifique et le monde industriel. La communauté scientifique a oeuvré pour la simplification de certains problèmes issus du monde industriel vers des modèles d’optimisation combinatoire. Parmi ces problèmes, on peut trouver des problèmes appartenant à la famille du problème du sac à dos (knapsack). Dans cette thèse, nous considérons une variante du problème du sac à dos : le problème du sac à dos avec des contraintes disjonctives (Knapsack with Disjunctive Constraints). En raison de la difficulté de cette problématique, nous nous sommes intéressés au développement de méthodes heuristiques produisant des solutions de bonne qualité en un temps de calcul modéré. Nos travaux de recherche s’appuient sur le principe de la recherche par voisinage. Bien que cette façon de faire nous conduise vers des solutions approchées,leur utilisation ainsi que les résultats que nous avons obtenus restent intéressants tout en gardant un temps d’exécution raisonnable. Afin de résoudre des instances de grande taille pour la problématique étudiée, nous avons proposé des méthodes séquentielles et parallèles. Ces deux techniques de résolution sont basées sur la recherche par voisinage. Dans un premier temps, une première méthode de recherche par voisinage aléatoire a été proposée. Elle s’appuie sur la combinaison de deux procédures : une première procédure qui cherche à construire une série de solutions partielles et une deuxième procédure qui complète chacune des solutions partielles courantes par une exploration de son voisinage. Ensuite, une deuxième méthode adaptative a été mise en place. Elle s’appuie sur un système d’optimisation par colonie de fourmis pour simuler une recherche guidée et une procédure de descente pour explorer au mieux les solutions produites au cours du processus de recherche. Finalement, une troisième méthode a été élaborée dans le but de faire évoluer la performance des méthodes de recherche par voisinage. Dans cette partie de nos travaux de recherche, nous avons proposé une recherche par voisinage aléatoire parallèle. Nous nous appuyés sur l’exploration simultanée de différents (sous) espaces de recherche par différents processeurs, où chaque processeur adopte sa propre stratégie aléatoire pour construire ses propres voisinages en fonction de ses informations internes récoltées / Combinatorial optimization problems are of high interest both for the scientific world and for the industrial world. The research community has simplified many practical situations as combinatorial optimization problems. Among these problems, we can find some problems belonging to the knapsack family. This thesis considers a particular problem belonging to the knapsack family, known as the disjunctively constrained knapsack problem. Because of the difficulty of this problem, we are searching for approximate solution techniques with fast solution times for its large scale instances. A promising way to solve the disjunctively constrained knapsack problem is to consider some techniques based upon the principle of neighborhood search. Although such techniques produce approximate solution methods, they allow us to present fast algorithms that yield interesting solutions within a short average running time. In order to tackle large scale instances of the disjunctively constrained knapsack problem, we present sequential and parallel algorithms based upon neighborhood search techniques. The first algorithm can be viewed as a random neighborhood search method. This algorithm uses a combination of neighborhood search techniques in order to randomly explore a series of sub-solution spaces, where each subspace is characterized by a neighborhood of a local optimum. The second algorithm is an adaptive neighborhood search that guides the search process in the feasible solution space towards high quality solutions. This algorithm uses an ant colony optimization system to simulate the guided search. The third andlast algorithm is a parallel random neighborhood search method which exploits the parallelism for exploring simultaneously different sub-solution spaces by several processors. Each processor adopts its own random strategy to yield its own neighborhoods according to its internal information

Advanced connected and automated vehicle technologies offer new opportunities for highway traffic smoothing by optimizing automated vehicle trajectories. As one of the pioneering attempts, this study proposes an efficient trajectory optimization algorithm that can simultaneously improve a range of performance measures for a platoon of vehicles on a signalized highway section. This optimization is centered at a novel shooting heuristic (SH) for trajectory construction that considers realistic constraints including vehicle kinematic limits, traffic arrival patterns, carollowing safety, and signal operations. SH has a very parsimonious structure (e.g., only four acceleration parameters) and a very small computational complexity. Therefore, it is suitable for real-time applications when relevant technologies are in place in the near future. This study lays a solid foundation for devising holistic cooperative control strategies on a general transportation network with emerging technologies.

Metamemory is a person’s knowledge about their own memory. Metamemory judgments are sometimes accurate and sometimes not. Eakin (2005) found a dissociation between metamemory predictions and memory performance under conditions of retroactive interference and attributed this dissociation to the accessibility heuristic. This study investigated whether the accessibility heuristic would be used to make metamemory predictions in the more complex context of the eyewitness memory paradigm. The results indicate that the accessibility heuristic was used to make metamamory predictions. Memory performance was better for control than misled critical items, but people predicted they would perform equally well in both conditions. It appears that in the less austere context of the eyewitness memory paradigm, the amount of information accessible for control and misled items was equal, and therefore, metamemory judgments were equal for control and misled items.