Coordenação de Aperfeiçoamento de Pessoal de Nível Superior / The Crop Livestock Integration System (ILP) is based on diversification, rotation, intercropping and/or succession of agriculture and livestock activities within the rural property. The aim of the study was to evaluated if different grazing intensities applied in the ILP change soil fauna diversity and activity, thus identifying soil quality indicators in soils under ILP systems. The experimental area consisted of 14 plots, where treatments are based on four grazing intensities (10, 20, 30 and 40 cm) regulated from the pasture height and distributed in an experimental design of randomized blocks. It was used three replicates, and two areas of ungrazed blocks as reference areas. Sampling was performed by five points to mites and nine points for the eartworms in each plot, pre-set by the use locations on a grid, with the same measurements, regardless of the size of plots, after grazing of cattle and after soybean harvest. Funnel Berleze-Tullgren Modified was assessed to verify the presence of edaphic mites. The Tropical Soil Biology and Fertility methodology (TSBF) was carried out to collected earthworms. The total mites density is favored by the 20cm height in after grazing of cattle in ILP system. In the assessment, there is a greater diversity of mites families (Upododidae, Pachilaelapidae, Nothridae, Phthiracaridae, Galumnidae, Oribatidae, Trombiculidae, Cunaxidae four unidentified) after soybean cultivation. The presence of earthworms were highest after soybean cultivation and of in 10-20cm deep layer on Integration Livestock System. The presence of two families of Glossoscolecidae and Ocnerodrilidae worms were verified in ILP area, with different grazing intensities. Moreover, the first occurrence of the species Eukerria michaelsen was recorded in the Rio Grande do Sul state, Brazil. / O sistema de Integração Lavoura Pecuária (ILP) é baseado na diversificação, rotação, consorciação e/ou sucessão das atividades de agricultura e de pecuária dentro da propriedade rural. Este trabalho tem por objetivo estudar influencia de diferentes intensidades de pastejo aplicadas na ILP, após colheita da soja e após saído dos bovinos, na ocorrência de ácaros edáficos e minhocas. A área experimental foi composta de 14 parcelas, onde os tratamentos baseiam-se em quatro intensidades de pastejo (10, 20, 30 e 40 cm) reguladas a partir da altura da pastagem. As parcelas são distribuídas em um delineamento experimental de blocos ao acaso, com três repetições, e duas áreas entre blocos sem pastejo, como áreas de referência. A amostragem foi realizada em cinco pontos para os ácaros e nove pontos para as minhocas em cada parcela, em locais pré-estabelecidos pelo uso de um grid, com as mesmas medidas, independente do tamanho da parcela, após o pastejo dos bovinos e após a colheita da soja. Para avaliação da presença dos ácaros edáficos utilizou-se Funil de Berleze-Tullgren Modificado. Para as minhocas foram coletadas por meio da metodologia do Tropical Soil Biology and Fertility (TSBF). A densidade total de ácaros é favorecida pela altura de 20cm no após pastejo dos bovinos no sistema ILP. Na avaliação, após o cultivo de soja, há uma maior diversidade de famílias de ácaros (Upododidae, Pachilaelapidae, Nothridae, Phthiracaridae, Galumnidae, Oribatidae, Trombiculidae, Cunaxidae e quatro morfotipos não identificados). Para a presença de minhocas há uma maior abundância após cultivo da soja na profundidade de 10-20cm no sistema estudado. Estas minhocas foram caracterizadas sendo de duas famílias, Glossoscolecidae e Ocnerodrilidae. Registro-se o primeiro relato da espécie Eukerria michaelsen, no estado do Rio Grande do Sul, Brasil.

Submitted by Valquíria Barbieri (kikibarbi@hotmail.com) on 2018-03-16T21:00:03Z No. of bitstreams: 1 _DISS_2014_Antonio Carlos Buchelt.pdf: 1033541 bytes, checksum: 7ca76020e7562b21c265b815f1f2564e (MD5) / Approved for entry into archive by Jordan (jordanbiblio@gmail.com) on 2018-04-10T16:54:15Z (GMT) No. of bitstreams: 1 _DISS_2014_Antonio Carlos Buchelt.pdf: 1033541 bytes, checksum: 7ca76020e7562b21c265b815f1f2564e (MD5) / Made available in DSpace on 2018-04-10T16:54:15Z (GMT). No. of bitstreams: 1 _DISS_2014_Antonio Carlos Buchelt.pdf: 1033541 bytes, checksum: 7ca76020e7562b21c265b815f1f2564e (MD5) Previous issue date: 2014-07-04 / CAPES / A pastagem degradada e a baixa fertilidade do solo são motivos de baixa produção da pecuária leiteira. O objetivo deste trabalho foi avaliar a fertilidade do solo em sistemas integrados de produção de leite, na região de transição de Cerrado para Floresta Amazônica. O experimento foi conduzido na Embrapa Agrossilvipastoril, localizada no município de Sinop – MT. O solo característico da região é o Latosssolo Vermelho-Amarelo distrófico. O experimento foi implantado a campo em 2011, sendo conduzido em uma área de 40 ha sendo divididos em quatro sistemas de cultivo, cada um com 10 ha (lavoura de 2° ano, pastagem de 1° ano, pastagem de 2° ano e lavoura de 1°ano) que foram subdivididos em três tratamentos (A: integração Lavoura-Pecuária, B: integração Lavoura-Pecuária-Floresta, com renques duplo de eucalipto nas laterais e C: integração Lavoura-Pecuária-Floresta, com renques triplos de eucalipto). As coletas das amostras para a análise química foram realizadas, nos quatros sistemas de cultivo e em pontos variáveis para cada um dos tratamentos com três repetições. As análises realizadas foram pH em água, fósforo, potássio, matéria orgânica, cálcio, magnésio e hidrogênio + alumínio, onde a partir desse valores foram determinados soma de bases, capacidade de troca de cátions a pH 7 e saturação por bases. Os dados foram analisados com auxílio do programa SAS® a 10 % de significância. No tratamento A, todas as variáveis com exceção do H+Al apresentaram diferença em algum dos sistemas de cultivo. Já os tratamentos B apresentaram diferença em todas as variáveis em algum dos sistemas de cultivo. No tratamento B o K foi o nutriente que apresentou diferença em todos os sistemas. Já no C a MO não diferiu em nenhum sistema de cultivo, no tratamento C e as concentrações médias variaram de 26,8-29,8 g kg-1. Em relação aos sistemas de cultivos o pH variou de 4,8 na mata a 5,5 no pousio, os cultivos apresentaram valores entre 5,2-5,4. As concentrações de K, Al+H e V, foram influenciados pelos sistemas de cultivos. A mata apresentou baixos valores com exceção do Al+H. As concentrações de K apresentaram grande variabilidade entre os sistemas de cultivos em todos os tratamentos. As concentrações de fósforo e potássio apresentaram variabilidade nos sistemas de cultivos avaliados, sendo que o potássio apresentou diferenças em todos os sistemas para os tratamentos A e B. A saturação por base obteve um aumento em todos os sistemas de cultivos e na área de pousio, após a ação antrópica. / The degraded pasture and low soil fertility are reasons for low production of dairy cattle. The objective of this study was to evaluate soil fertility in integrated systems of milk production in the Cerrado region of transition to the Amazon rainforest. The experiment was conducted at Embrapa agrosilvopastoral, located at Sinop - MT. The soil in the area is the Red-Yellow Latosssolo dystrophic. The experiment was deployed into the field in 2011, being conducted in an area of 40 ha were divided into four cultivation systems, each with 10 ha (2nd year crop, pasture 1st year, 2nd year of grazing and farming 1 ° year) who were divided into three treatments (a: Crop-Livestock integration, B: Integration of Agriculture-Forest, with double rows of eucalyptus on the sides and C: integration Agriculture-Forest, with triple rows of eucalyptus ). The collection of samples for chemical analysis were performed in four cropping systems and variables for each treatment with three replications points. The analyzes were pH, phosphorus, potassium, organic matter, calcium, magnesium and hydrogen + aluminum, where the values were determined from this sum of bases, cation exchange capacity at pH 7 and base saturation. Data were analyzed using the SAS ® 10% significance. In treatment A, all variables except for the H + Al showed differences in some farming systems. Whereas the B treatments showed differences in all variables in some farming systems. In treatment B, K was the nutrient that was different on all systems. You C MO did not differ in any culture system, treatment C and the average concentrations ranged from 26.8 to 29.8 g kg-1. Regarding cropping systems the pH ranged from 4.8 to 5.5 in the forest fallow crops varied from 5.2-5.4. The concentrations of K, Al + H and V, were influenced by cropping systems. The forest showed low values except for Al + H. The K concentrations varied considerably among cropping systems in all treatments. The concentrations of phosphorus and potassium showed variability in crop systems evaluated, and potassium showed differences in all systems for treatments A and B. The base saturation obtained an increase in all cropping systems and fallow, after human action.

Title: Project portfolio optimization with time and resources Author: Bc. Tomáš Huml Department: Department of Theoretical Computer Science and Mathematical Logic Supervisor: Doc. RNDr. Roman Barták, Ph.D Abstract: Traditional project portfolio optimization deals with static projects that are not evolving in time. The focus of this diploma thesis is on projects that are spread in time, typically such projects consists of a sequence (or other partially ordered structure) of actions that require some resources (money, people, etc.) for realization. Then the project portfolio optimization deals with selecting a subset of projects according to given time and space (resource) restrictions and optimizing certain criteria such as overall profit. This problem is very close to oversubscribed scheduling where the most profitable subset of orders is being scheduled. Hence scheduling techniques will be the main inspiration for solving this new type of problems. Lots of modelling algorithms for optimal portfolio selection are proposed in this diploma thesis and several of them are implemented in a program which is part of this thesis as well. Keywords: portfolio optimization, integer linear programming (ILP), workflow optimization, project interdependencies

Cette thèse contribue au contrôle de la myopie dans les systèmes flexibles de production (SFP). La myopie apparaît lorsque des entités décisionnelles prennent des décisions locales à partir d’une quantité d’information limitée. Cette prise de décision permet de réagir rapidement aux aléas mais induit une performance globale non optimale. Ainsi, ce phénomène doit être contrôlé afin d’obtenir des architectures de pilotage plus performantes. Après une étude du phénomène de myopie dans d’autres domaines, nous définissons la myopie dans les SFP. Un état de l’art sur les différents types d’architectures permet de retenir les architectures de pilotage hybride mixant hiérarchie et hétérarchie pour contrôler la myopie. Une typologie des architectures de pilotage hybride est ensuite réalisée avant qu’une nouvelle architecture ne soit proposée : ORCA. Après avoir été présentée, ORCA est déclinée au pilotage des SFP (ORCA-FMS). ORCA-FMS combine deux approches: un modèle linéaire (ILP) et une approche par champs de potentiel. ORCA-FMS est ensuite appliquée au cas d’étude de la cellule flexible de l’AIP PRIMECA de Valenciennes. Premièrement un modèle de simulation, le plus proche possible du cas d’étude réel est présenté. Il permet d’éprouver l’architecture dans l’environnement de simulation NetLogo. Deuxièmement, afin de valider la pertinence des comportements observés en simulation, l’architecture est mise en œuvre sur la cellule réelle à l’aide du concept de produit actif. Les équipements industriels utilisés pour cette mise en œuvre, le protocole expérimental, ainsi que les résultats obtenus sont détaillés et discutés. / This thesis deals with the control of myopia in Flexible Manufacturing Systems (FMS). Myopia arises when decisional entities take local decisions using limited amount of information. This decision making targets a fast reactivity under perturbations but compromises the overall performance. Thus this phenomenon should be controlled to obtain more efficient control architectures. After a presenting the related literature in myopia in other domains, myopic behavior in FMS is defined. An analysis of state-of-the-art regarding different types of control architectures determined that hybrid control architectures, mixing hierarchy and heterarchy, are the best option to control myopia. Therefore, a thorough study on hybrid control architectures is presented. Afterwards, a new architecture is proposed: ORCA. ORCA first described and then applied to FMS control (ORCA-FMS). ORCA-FMS combines two approaches: a linear model (ILP) and a potential fields approach. ORCA-FMS is then applied to the case study of the flexible cell of Valenciennes’ AIP PRIMECA. First, a simulation model, as close as possible to the real case study is presented. It allows testing the architecture in the simulation environment NetLogo. Then, to validate the behaviors observed in simulation, the architecture is implemented on the real cell using the active product concept. The industrial equipment used for the implementation, the experimental protocol and the results are detailed and discussed.

Interruptions to juvenile detainees' education often delay their progress toward high school completion. Implementing an individualized learning plan (ILP) has been suggested as a solution to this problem. The purpose of this case study was to explore how ILPs facilitate attainment of graduation among incarcerated youth. The study was guided by the Washington State legislative framework for individualized learning plans and the efforts of the Office of Juvenile Justice and Delinquency to increase educational opportunities. Three research questions were formulated to explore perceptions of local juvenile detention center educators and administrators regarding their experiences in the development, implementation, and effectiveness of ILPs. Qualitative data were collected from interviews with 5 detention center educators, field observations, and document reviews. The examination of the data through the coding process using a matrix enabled descriptions and themes to emerge. Results indicated that ILPs are collaboratively developed and implemented in the detention center, there is a need for professional development related to ILPs, and ILPs have impacted students' education completion. Results from this study may influence social change by supporting development of ILPs which can result in a higher graduation rate and a reduction in recidivism among the incarcerated youth population.

The growing interest that multimedia processing has experimented during the last decade is motivating processor designers to reconsider which execution paradigms are the most appropriate for general-purpose processors. On the other hand, as the size of transistors decreases, power dissipation has become a relevant limitation to increases in the frequency of operation. Thus, the efficient exploitation of the different sources of parallelism is a key point to investigate in order to sustain the performance improvement rate of processors and face the growing requirements of future multimedia applications. We belief that a promising option arises from the combination of the Very Long Instruction Word (VLIW) and the vector processing paradigms together with other ways of exploiting coarser grain parallelism, such as Chip MultiProcessing (CMP). As part of this thesis, we analyze the problem of memory disambiguation in multimedia applications, as it represents a serious restriction for exploiting Instruction Level Parallelism (ILP) in VLIW architectures. We state that the real handicap for memory disambiguation in multimedia is the extensive use of pointers and indirect references usually found in those codes, together with the limited static information available to the compiler on certain occasions. Based on the observation that the input and output multimedia streams are commonly disjointed memory regions, we propose and implement a memory disambiguation technique that dynamically analyzes the region domain of every load and store before entering a loop, evaluates whether or not the full loop is disambiguated and executes the corresponding loop version. This mechanism does not require any additional hardware or instructions and has negligible effects over compilation time and code size. The performance achieved is comparable to that of advanced interprocedural pointer analysis techniques, with considerably less software complexity. We also demonstrate that both techniques can be combined to improve performance.In order to deal with the inherent Data Level Parallelism (DLP) of multimedia kernels without disrupting the existing core designs, major processor manufacturers have chosen to include MMX-like µSIMD extensions. By analyzing the scalability of the DLP and non-DLP regions of code separately in VLIW processors with µSIMD extensions, we observe that the performance of the overall application is dominated by the performance of the non-DLP regions, which in fact exhibit only modest amounts of ILP. As a result, the performance achieved by very wide issue configurations does not compensate for the related cost. To exploit the DLP of the vector regions in a more efficient way, we propose enhancing the µSIMD -VLIW core with conventional vector processing capabilities. The combination of conventional and sub-word level vector processing results in a 2-dimensional extension that combines the best of each one, including a reduction in the number of operations, lower fetch bandwidth requirements, simplicity of the control unit, power efficiency, scalability, and support for multimedia specific features such as saturation or reduction. This enhancement has a minimal impact on the VLIW core and reaches more parallelism than wider issue µSIMD implementations at a lower cost. Similar proposals have been successfully evaluated for superscalar cores. In this thesis, we demonstrate that 2-dimensional Vector-µSIMD extensions are also effective with static scheduling, allowing for high-performance cost-effective implementations.

Real-world learning tasks present a range of issues for learning systems. Learning tasks can be complex and the training data noisy. When operating as part of a larger system, there may be limitations on available memory and computational resources. Learners may also be required to provide results from a stream. This thesis investigates the problem of incremental, relational learning from imperfect data with constrained time and memory resources. The learning process involves incremental update of a theory when an example is presented that contradicts the theory. Contradictions occur if there is an incorrect theory or noisy data. The learner cannot discriminate between the two possibilities, so both are considered and the better possibility used. Additionally, all changes to the theory must have support from multiple examples. These two principles allow learning from imperfect data. The Minimum Description Length principle is used for selection between possible worlds and determining appropriate levels of additional justification. A new encoding scheme allows the use of MDL within the framework of Inductive Logic Programming. Examples must be stored to provide additional justification for revisions without violating resource requirements. A new algorithm determines when to discard examples, minimising total usage while ensuring sufficient storage for justifications. Searching for revisions is the most computationally expensive part of the process, yet not all searches are successful. Another new algorithm uses a notion of theory stability as a guide to occasionally disallow entire searches to reduce overall time. The approach has been implemented as a learner called NILE. Empirical tests include two challenging domains where this type of learner acts as one component of a larger task. The first of these involves recognition of behavior activation conditions in another agent as part of an opponent modeling task. The second, more challenging task is learning to identify objects in visual images by recognising relationships between image features. These experiments highlight NILE'S strengths and limitations as well as providing new n domains for future work in ILP.

Η παρούσα εργασία έχει ως στόχο να παρουσιάσει έναν παράλληλο αλγόριθμο Εξόρυξης δεδομένων που μπορεί να εφαρμοστεί σε μεγάλες βάσεις δεδομένων με χρήση Επαγωγικού Λογικού Προγραμματισμού. Η κεντρική υπόθεση αυτής της εργασίας είναι πως είναι απαραίτητο αλλά και δυνατό να υιοθετηθούν παράλληλοι αλγόριθμοι στη διαδικασία της Εξόρυξης δεδομένων. Αποδεικνύεται ότι ο παραλληλισμός μπορεί να εφαρμοστεί αποδοτικά στον Επαγωγικό Λογικό Προγραμματισμό (Inductive Logic Programming - ILP). Η ολοκληρωμένη αναπαράσταση της γνώσης και η εξαιρετική ενσωμάτωση της υποκείμενης γνώσης του ILP αποδεικνύεται πολύ σημαντική μεταξύ των αλγορίθμων Εξόρυξης δεδομένων. Το Κεφάλαιο 1 αποτελεί μια εισαγωγή στην εξόρυξη δεδομένων. Παρουσιάζονται οι βασικές εργασίες που λαμβάνουν χώρα κατά την εξόρυξη δεδομένων καθώς και μερικοί από τους συνηθισμένους αλγορίθμους εξόρυξης δεδομένων. Στη συνέχεια, γίνεται μια εισαγωγή στη σχεσιακή εξόρυξη δεδομένων και δίνονται τα βασικά χαρακτηριστικά αλγορίθμων σχεσιακής εξόρυξης δεδομένων. Το Κεφάλαιο 2 αυτής της διπλωματικής εργασίας παρουσιάζει τη θεωρία του Eπαγωγικού Λογικού Προγραμματισμού (ΕΠΛ - ILP), μερικές δημοφιλείς μεθόδους ILP καθώς και τον τρόπο με τον οποίο εφαρμόζεται ο ILP για να επιλύσει προβλήματα που συναντώνται σε συστήματα βάσεων δεδομένων Παρουσιάζεται αναλυτικά μια συγκεκριμένη προσέγγιση στον ILP – η Mode Directed Inverse Entailment [60], καθώς αποτελεί τη βάση για την παράλληλη έκδοση της Progol που θα παρουσιαστεί σε επόμενο κεφάλαιο. Το Κεφάλαιο 3 είναι αφιερωμένο στην παρουσίαση διαφόρων συστημάτων ILP και την εφαρμογή τους στην Εξόρυξη δεδομένων. Στη συνέχεια του κεφαλαίου παρουσιάζεται ένας σειριακός αλγόριθμος ILP Εξόρυξης δεδομένων και η ανάλυση κόστους του. Γίνεται μια εισαγωγή στο Bulk Synchronous Parallelism (BSP) [46]. Το Κεφάλαιο 4, είναι μια σύνοψη της παράλληλης Επαγωγικής Λογικής. Στην αρχή αναφέρονται οι βασικές αρχές του παράλληλου επαγωγικού μοντέλου και στη συνέχεια ακολουθεί μια λεπτομερής περιγραφή. Εμφανίζονται τα προβλήματα μιας τέτοιας προσέγγισης και παρουσιάζεται μια ανάλυση κόστους. Τμήμα του Κεφαλαίου 4 είναι αφιερωμένο σε έναν παράλληλο αλγόριθμο ILP που χρησιμοποιεί το μοντέλο BSP. Για να υποστηριχθεί ο παράλληλος αλγόριθμος του Κεφαλαίου 4, το Κεφάλαιο 5 παρουσιάζει μερικά βασικά χαρακτηριστικά ενός παράλληλου συστήματος ILP – της παράλληλης Progol. Το σύστημα αυτό δημιουργήθηκε με βάση το μοντέλο BSP και βασίζεται στην Progol που υλοποιήθηκε (σε γλώσσα C) από το Muggleton. Παρέχονται πολλά παραδείγματα και επεξηγείται το φαινόμενο της υπεργραμμικής επιτάχυνσης (super-linear speedup). Στο Κεφάλαιο 6 έχουν συμπεριληφθεί τα αποτελέσματα των δοκιμών του αλγορίθμου του Κεφαλαίου 4 σε τρεις διαφορετικές περιπτώσεις προβλημάτων, στην ταξινόμηση των ζώων, στην εκμάθηση των κανόνων του σκάκι και στο πρόβλημα τερματισμού σε μια παρτίδα σκάκι. Το Κεφάλαιο 7 συνοψίζει τα αποτελέσματα της μεταπτυχιακής εργασίας. / -

Interruptions to juvenile detainees' education often delay their progress toward high school completion. Implementing an individualized learning plan (ILP) has been suggested as a solution to this problem. The purpose of this case study was to explore how ILPs facilitate attainment of graduation among incarcerated youth. The study was guided by the Washington State legislative framework for individualized learning plans and the efforts of the Office of Juvenile Justice and Delinquency to increase educational opportunities. Three research questions were formulated to explore perceptions of local juvenile detention center educators and administrators regarding their experiences in the development, implementation, and effectiveness of ILPs. Qualitative data were collected from interviews with 5 detention center educators, field observations, and document reviews. The examination of the data through the coding process using a matrix enabled descriptions and themes to emerge. Results indicated that ILPs are collaboratively developed and implemented in the detention center, there is a need for professional development related to ILPs, and ILPs have impacted students' education completion. Results from this study may influence social change by supporting development of ILPs which can result in a higher graduation rate and a reduction in recidivism among the incarcerated youth population.

Cette thèse propose un schéma entièrement distribué de routage résilient pour les réseaux de switches. Dans cette approche une panne est traitée localement ; ainsi les autres nœuds du réseau n’ont pas besoin de prendre des mesures spécifiques. Contrairement aux systèmes classiques de routage IP qui se base uniquement sur la destination, le routage est fait en se basant sur l’arc de l’entrée et la destination. La contrainte qui en résulte est que les deux flux à la même destination entrant dans un nœud par un arc commun doivent fusionner après cet arc. Nous montrons que sous des faibles hypothèses (en supposant que le réseau est symétrique et bi-connexes au sens des liens), il existe un schéma de routage résilient pour toutes les situations de pannes simples de liens. Nous modélisons le problème de dimensionnement par un programme linéaire en nombres entiers qui peut être résolu exactement pour des réseaux de taille modeste. Nous proposons également plusieurs heuristiques pour traiter des instances de grande taille. Notre méthode généralise les méthodes de la littérature, en particulier celles de Xi et al. et de Nelakuditi et al. qui ont proposé des approches similaires. Notre approche permet d’obtenir de meilleurs résultats sur de nombreuses instances de test. Nous avons également étudié l’existence d’un schéma de routage résilient pour situation de panne de nœud dans le réseau basée sur les switches. Nous montrons que dans le cas des pannes de plusieurs liens, la condition de connexité n’est pas suffisante. / This thesis presents a fully distributed resilient routing scheme for switch-based networks. A failure is treated locally, so other nodes in the network do not need to undertakespecial actions. In contrast to conventional IP routing schemes, each node routesthe traffic on the basis of the entering arc and of the destination. The resulting constraintis that two flows to the same destination entering in a node by a common archave to merge after this arc. We show that this is sufficient for dealing with all singlelink failure situations, assuming that the network is symmetric and two-link connected.We model the dimensioning problem with an Integer Linear Program which can besolved exactly for small networks. We also propose several heuristics for larger networks.Our method generalizes the methods of Xi et Chao and Li and of Nelakuditiet al. who have proposed similar schemes in the context of IP. Our methods are moreefficient than previous ones. We have also studied the existence of a resilient routingscheme for single node failure situation in switch-based network. We study also thecase of multi-link failure situations and show that requiring the network to be connectedafter any failure does not guarantee the existence of a resilent routing schemeas described above.