Les travaux présentés dans cette thèse visent à proposer des méthodes permettant de résoudre un problème de Transport à la Demande multicritère. Le premier travail réalisé dans cette thèse est l'étude d'un problème de Dial-a-Ride (DARP) statique multicritère. Trois critères qui peuvent être conflictuels ont été définis : le premier consiste à minimiser le coût de transport, le deuxième critère consiste à minimiser l'insatisfaction des passagers et enfin le troisième critère consiste à minimiser la quantité de CO2 émise par l'ensemble des véhicules. Nous avons développé une méthode évolutionnaire NSGA-II pour chercher un ensemble approximatif d'optimas de Pareto. Le second travail réalisé est l'étude d'un problème d'Optimal Timing dans une tournée. Ce problème consiste à calculer les dates de début de service optimales des points d'arrêts d'une tournée afin de minimiser l'insatisfaction des passagers. Le dernier travail de cette thèse a porté sur l'étude d'un problème de Transport à la Demande dynamique dans lequel de nouvelles requêtes à traiter arrivent en cours de journée. Deux méthodes ont été proposées pour résoudre ce problème : la première est une heuristique d'insertion rapide et la seconde est une méthode arborescente tronquée connue sous le nom de Recovering Beam Search. / The work presented in this thesis aims to propose methods to solve a multicriteria dial-a-ride problem (DARP). Three objective functions that have to be optimized in order to measure the potential efficiency of the DARP solution on different aspects : the cost for the transportation operator, the quality of service for users and the impact on the environment. The first work in this thesis is the study of static DARP for which a NSGA-II algorithm is developped to identify a good approximation of the Pareto optimal set. The second work deals with an optimal timing algorithm which computes pickup and delivery dates when the requests are sequenced on the vehicles, the objective is to minimize the total customer' dissatisfaction. The last problem studied in this thesis aims to solve the dynamic version of DARP for which two methods are proposed. The first one is a fast insertion heuristic based on an attractive index. However, the second methode uses a recovering beam search heuristic which unlike the insertion heuristic allows to modify the structure of the routes previously scheduled in order to schedule the new requests.

Submitted by Marlene Santos (marlene.bc.ufg@gmail.com) on 2014-12-16T20:28:04Z No. of bitstreams: 2 Dissertação - Carlos Antônio Campos Jorge - 2014.pdf: 703425 bytes, checksum: 664e77c2f8e857788e0128256d76d4b7 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Approved for entry into archive by Luciana Ferreira (lucgeral@gmail.com) on 2014-12-22T10:40:49Z (GMT) No. of bitstreams: 2 Dissertação - Carlos Antônio Campos Jorge - 2014.pdf: 703425 bytes, checksum: 664e77c2f8e857788e0128256d76d4b7 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) / Made available in DSpace on 2014-12-22T10:40:49Z (GMT). No. of bitstreams: 2 Dissertação - Carlos Antônio Campos Jorge - 2014.pdf: 703425 bytes, checksum: 664e77c2f8e857788e0128256d76d4b7 (MD5) license_rdf: 23148 bytes, checksum: 9da0b6dfac957114c6a7714714b86306 (MD5) Previous issue date: 2014-04-08 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / This work proposes the use of a multi-objective evolutionary algorithm that makes use of subsets stored in a data structure called table in which the best individuals from each objective considered are preserved. This approach is compared in this work with the traditional mono-objective evolutionary algorithm (GA), classical algorithms (PLS and SPA) and another classic multi-objective algorithm (NSGA-II). As a case study, a multivariate calibration problem is presented which involves the prediction of protein concentration in samples of whole wheat from the spectrophotometric measurements. The results showed that the proposed formulation has a smaller prediction error when compared to the mono-objective formulation and with a lower number of variables. Finally,astudyofnoisesensitivityobtainedbythemulti-objectiveformulationshoweda better resultwhen compared tothe other classical algorithmforvariable selection. / Este trabalho propõe o uso de algoritmo multi-objetivo evolutivo que faz uso de subconjuntos armazenados em uma estrutura de dados chamada tabela em que os melhores indivíduos de cada objetivo são preservadas. Esta abordagem é comparada neste trabalho com o algoritmo evolutivo tradicional mono-objetivo e outros algoritmos clássicos (MONO-GA-MLR, PLS, APS-MLR) e com o algoritmo multi-objetivo clássico NSGAII-MLR.Comoestudodecaso,oproblemadecalibraçãomultivariadaenvolveaprevisão daconcentraçãodeproteínasemamostrasdetrigoapartirdasmediçõesespectrofotométricas. Os resultados mostraram que a formulação proposta seleciona um número menor de variáveis e apresenta um erro de predição menor quando comparada com o algoritmo evolutivo mono-objetivo. Quando comparado com os algoritmos clássicos PLS e APSMLR e com o algoritmo multi-objetivo clássico NSGA-II-MLR, o algoritmo proposto apresenta um erro de predição menor, porém com um número maior de variáveis selecionadas. Finalmente, um estudo de sensibilidade à ruído foi realizado. A solução obtida pela formulação proposta apresentou melhores resultados quando comparado com o algoritmo mono-objetivo e NSGA-II-MLR e desempenho similar à solução obtida com o SPA-MLR.

This thesis is focused on working up evolutionals and genetics algorithms issues Especially for multiobjective algorithms VEGA, SPEA and NSGA – II. Thereinafter one of FrameWork working with genetics algorithms namely WWW NIMBUS. From this mentioned algorithms was selected VEGA algorithm for implementation in JAVA to preselected problem. Thereby problem is choice thick columns of profile according to predetermined criteria. Selected algorithm works on division of population into several groups and each group evaluates the resulting fitness function. Here is a sample implementation of this algorithm. Furthermore there is a example of working with FrameWork. In the next section are compared the results of generated progam with results that were obtained by FrameWork WWW NIMBUS. As for VEGA, and the Nimbus there are shown different results. The VEGA is presented also the development of individual fitness functions. Also, there are shown graphs, that can be obtained from NIMBUS. At the end of work is introduced the comparation of the results ane propose possible improvements.