Search

Spatial and temporal decomposition of aggregated mobility flows is nowadays a commonly addressed issue, but a trip-purpose decomposition of mobility flows is a more challenging topic, which requires more sensitive analysis such as heterogeneous data fusion. In this paper, we study the relation between land use and mobility purposes. We propose a model that dynamically decomposes mobility flows into six mobility purposes. To this end, we use a national transportation database that surveyed more than 35,000 individuals and a national ground description database that identifies six distinct ground types. Based on these two types of data, we dynamically solve several overdetermined systems of linear equations from a training set and we infer the travel purposes. Our experimental results demonstrate that our model effectively predicts the purposes of mobility from the land use. Furthermore, our model shows great results compared with a reference supervised learning decomposition.

De nombreux problèmes réels nécessitent d'optimiser plusieurs objectifs, souvent en conflit les uns avec les autres. Cette famille de problèmes appelés multi-objectifs (MOP) n'admet pas une solution unique mais un ensemble de solutions de compromis dit non-dominé. Les algorithmes multi-objectifs visent à construire les solutions se rapprochant au mieux de l'ensemble Pareto optimal, ensemble qui domine l'espace de recherche. L'hybridation de plusieurs métaheuristiques pour résoudre les MOP s'est particulièrement développée ses dernières années. Parmi les algorithmes les plus performants figurent les approches mémétiques, hybridant recherches locales et algorithmes génétiques. L'algorithme MEMO que nous présentons repose sur une recherche locale basée sur une descente simple et sur l'algorithme génétique NSGA-II. Un des éléments les plus délicats à aborder est l'équilibre entre intensification (meilleures exploitation d'une solution) et diversification (meilleure exploration de l'espace de recherche). Différents composants d'un algorithme mémétique permettent d'agir tant sur la diversification que sur l'intensification. Ainsi par exemple, le croisement de deux solutions tend à fournir plus de diversité alors que la recherche locale tend à fournir une meilleure intensification. Mais cette tendance peut toutefois être modulée selon le croisement choisi ou la recherche locale. Nous proposons ici une étude détaillée de l'influence de cet équilibre entre intensification et diversification sur la qualité de l'ensemble des solutions non-dominées trouvées par MEMO. Le cas d'étude retenu est le problème de localisation de stations d'auto-partage électrique, modéliser sous forme d'un problème à trois objectifs. De plus, les résultats obtenus par l'algorithme MEMO sont comparés aux résultats obtenus avec les approches génétiques et locales seules. Il s'est révélé être particulièrement performant. Pour plus de détail sur l'algorithme et le problème de localisation de stations d'auto-partage : http://dl.acm.org/citation.cfm?id=2464627&dl=ACM&coll=DL&CFID=381215587&CFTOKEN=78886318

"In one-way carsharing systems, users can take a car from a station and return it to any other one. Available cars and free parking space at each station, play a major role in the success of these systems. Therefore, carsharing operators recruit employees to relocate cars between the stations to avoid rejection of users’ demands for taking cars or returning them. In this paper, we developed a greedy algorithm in order reduce the number of rejected users' demands, using the minimum number of employees. We used mobility data collected from an operational system to build the users’ demand matrices as input for our algorithm. For example, in a carsharing system which has 20 stations, 150 cars, through 1,374 trips made during one day and 556 expected rejected demands, results show that our algorithm is able to solve 55% of expected rejected demands using 5 employees and more than 80% of these rejected demands using 10 employees. We compared the algorithm with an exact Mixed Integer Linear Programming model solved using IBM ILOG CPLEX optimizer and made the proof of performance also, we evaluated the robustness of the two approaches with stochastic input data using different numbers of employees."

International audience; Multimedia traffic is constantly increasing and will soon dominate traffic flows in radio networks. The Multimedia Broadcast Multicast Service (MBMS) system provides efficient mechanisms for multimedia multicast services in mobile networks. We develop a flexible model to perform dynamic radio resource allocation for MBMS service by using metaheuristics approach. We conduct fitness landscape analysis to study the characteristics of the proposed problem, which helps us to select appropriate search strategy. Simulation results show that the proposed algorithm provides better performance than existing algorithms.

In this article, we present a method combining a genetic approach with a local search for multiobjective problems. It is an extension of algorithms for the single objective case, with specific mechanisms used to build the Pareto set. The performance of the proposed algorithm is illustrated by experimental results based on a real problem with three objectives. The problem is issued from electric carsharing service with a car manufacturer partner. Compared to the Multiobjective Pareto Local Search well known in the scientific literature, the proposed model aims to improve: the solutions quality and the set diversity.