Efficient asynchronous scheduling algorithms for cost-effective buffered crossbar switches

Buffered crossbar switches have recently attracted considerable attention as the next generation of high speed interconnects. They are a special type of crossbar switches with an exclusive buffer at each crosspoint of the crossbar. They demonstrate unique advantages over traditional unbuffered crossbar switches, such as high throughput, low latency, and asynchronous packet scheduling. However, since crosspoint buffers are expensive on-chip memories, it is desired that each crosspoint has only a small buffer. This dissertation proposes a series of practical algorithms and techniques for efficient packet scheduling for buffered crossbar switches. To reduce the hardware cost of such switches and make them scalable, we considered partially buffered crossbars, whose crosspoint buffers can be of an arbitrarily small size. Firstly, we introduced a hybrid scheme called Packet-mode Asynchronous Scheduling Algorithm (PASA) to schedule best effort traffic. PASA combines the features of both distributed and centralized scheduling algorithms and can directly handle variable length packets without Segmentation And Reassembly (SAR). We showed by theoretical analysis that it achieves 100% throughput for any admissible traffic in a crossbar with a speedup of two. Moreover, outputs in PASA have a large probability to avoid the more time-consuming centralized scheduling process, and thus make fast scheduling decisions. Secondly, we proposed the Fair Asynchronous Segment Scheduling (FASS) algorithm to handle guaranteed performance traffic with explicit flow rates. FASS reduces the crosspoint buffer size by dividing packets into shorter segments before transmission. It also provides tight constant performance guarantees by emulating the ideal Generalized Processor Sharing (GPS) model. Furthermore, FASS requires no speedup for the crossbar, lowering the hardware cost and improving the switch capacity. Thirdly, we presented a bandwidth allocation scheme called Queue Length Proportional (QLP) to apply FASS to best effort traffic. QLP dynamically obtains a feasible bandwidth allocation matrix based on the queue length information, and thus assists the crossbar switch to be more work-conserving. The feasibility and stability of QLP were proved, no matter whether the traffic distribution is uniform or non-uniform. Hence, based on bandwidth allocation of QLP, FASS can also achieve 100% throughput for best effort traffic in a crossbar without speedup.

Implementing multicore real-time scheduling algorithms based on task splitting using ada 2012

Multiprocessors, particularly in the form of multicores, are becoming standard building blocks for executing reliable software. But their use for applications with hard real-time requirements is non-trivial. Well-known realtime scheduling algorithms in the uniprocessor context (Rate-Monotonic [1] or Earliest-Deadline-First [1]) do not perform well on multiprocessors. For this reason the scientific community in the area of real-time systems has produced new algorithms specifically for multiprocessors. In the meanwhile, a proposal [2] exists for extending the Ada language with new basic constructs which can be used for implementing new algorithms for real-time scheduling; the family of task splitting algorithms is one of them which was emphasized in the proposal [2]. Consequently, assessing whether existing task splitting multiprocessor scheduling algorithms can be implemented with these constructs is paramount. In this paper we present a list of state-of-art task-splitting multiprocessor scheduling algorithms and, for each of them, we present detailed Ada code that uses the new constructs.

Real-time scheduling of embedded applications on multi-core platforms

For the past several decades, we have experienced the tremendous growth, in both scale and scope, of real-time embedded systems, thanks largely to the advances in IC technology. However, the traditional approach to get performance boost by increasing CPU frequency has been a way of past. Researchers from both industry and academia are turning their focus to multi-core architectures for continuous improvement of computing performance. In our research, we seek to develop efficient scheduling algorithms and analysis methods in the design of real-time embedded systems on multi-core platforms. Real-time systems are the ones with the response time as critical as the logical correctness of computational results. In addition, a variety of stringent constraints such as power/energy consumption, peak temperature and reliability are also imposed to these systems. Therefore, real-time scheduling plays a critical role in design of such computing systems at the system level. We started our research by addressing timing constraints for real-time applications on multi-core platforms, and developed both partitioned and semi-partitioned scheduling algorithms to schedule fixed priority, periodic, and hard real-time tasks on multi-core platforms. Then we extended our research by taking temperature constraints into consideration. We developed a closed-form solution to capture temperature dynamics for a given periodic voltage schedule on multi-core platforms, and also developed three methods to check the feasibility of a periodic real-time schedule under peak temperature constraint. We further extended our research by incorporating the power/energy constraint with thermal awareness into our research problem. We investigated the energy estimation problem on multi-core platforms, and developed a computation efficient method to calculate the energy consumption for a given voltage schedule on a multi-core platform. In this dissertation, we present our research in details and demonstrate the effectiveness and efficiency of our approaches with extensive experimental results.

'Explicit Learning: an Effort towards Human Scheduling Algorithms'

Abstract Scheduling problems are generally NP-hard combinatorial problems, and a lot of research has been done to solve these problems heuristically. However, most of the previous approaches are problem-specific and research into the development of a general scheduling algorithm is still in its infancy. Mimicking the natural evolutionary process of the survival of the fittest, Genetic Algorithms (GAs) have attracted much attention in solving difficult scheduling problems in recent years. Some obstacles exist when using GAs: there is no canonical mechanism to deal with constraints, which are commonly met in most real-world scheduling problems, and small changes to a solution are difficult. To overcome both difficulties, indirect approaches have been presented (in [1] and [2]) for nurse scheduling and driver scheduling, where GAs are used by mapping the solution space, and separate decoding routines then build solutions to the original problem. In our previous indirect GAs, learning is implicit and is restricted to the efficient adjustment of weights for a set of rules that are used to construct schedules. The major limitation of those approaches is that they learn in a non-human way: like most existing construction algorithms, once the best weight combination is found, the rules used in the construction process are fixed at each iteration. However, normally a long sequence of moves is needed to construct a schedule and using fixed rules at each move is thus unreasonable and not coherent with human learning processes. When a human scheduler is working, he normally builds a schedule step by step following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not completed yet, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this research we intend to design more human-like scheduling algorithms, by using ideas derived from Bayesian Optimization Algorithms (BOA) and Learning Classifier Systems (LCS) to implement explicit learning from past solutions. BOA can be applied to learn to identify good partial solutions and to complete them by building a Bayesian network of the joint distribution of solutions [3]. A Bayesian network is a directed acyclic graph with each node corresponding to one variable, and each variable corresponding to individual rule by which a schedule will be constructed step by step. The conditional probabilities are computed according to an initial set of promising solutions. Subsequently, each new instance for each node is generated by using the corresponding conditional probabilities, until values for all nodes have been generated. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the Bayesian network is updated again using the current set of good rule strings. The algorithm thereby tries to explicitly identify and mix promising building blocks. It should be noted that for most scheduling problems the structure of the network model is known and all the variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus learning can amount to 'counting' in the case of multinomial distributions. In the LCS approach, each rule has its strength showing its current usefulness in the system, and this strength is constantly assessed [4]. To implement sophisticated learning based on previous solutions, an improved LCS-based algorithm is designed, which consists of the following three steps. The initialization step is to assign each rule at each stage a constant initial strength. Then rules are selected by using the Roulette Wheel strategy. The next step is to reinforce the strengths of the rules used in the previous solution, keeping the strength of unused rules unchanged. The selection step is to select fitter rules for the next generation. It is envisaged that the LCS part of the algorithm will be used as a hill climber to the BOA algorithm. This is exciting and ambitious research, which might provide the stepping-stone for a new class of scheduling algorithms. Data sets from nurse scheduling and mall problems will be used as test-beds. It is envisaged that once the concept has been proven successful, it will be implemented into general scheduling algorithms. It is also hoped that this research will give some preliminary answers about how to include human-like learning into scheduling algorithms and may therefore be of interest to researchers and practitioners in areas of scheduling and evolutionary computation. References 1. Aickelin, U. and Dowsland, K. (2003) 'Indirect Genetic Algorithm for a Nurse Scheduling Problem', Computer & Operational Research (in print). 2. Li, J. and Kwan, R.S.K. (2003), 'Fuzzy Genetic Algorithm for Driver Scheduling', European Journal of Operational Research 147(2): 334-344. 3. Pelikan, M., Goldberg, D. and Cantu-Paz, E. (1999) 'BOA: The Bayesian Optimization Algorithm', IlliGAL Report No 99003, University of Illinois. 4. Wilson, S. (1994) 'ZCS: A Zeroth-level Classifier System', Evolutionary Computation 2(1), pp 1-18.

'Explicit Learning: an Effort towards Human Scheduling Algorithms'

Abstract Scheduling problems are generally NP-hard combinatorial problems, and a lot of research has been done to solve these problems heuristically. However, most of the previous approaches are problem-specific and research into the development of a general scheduling algorithm is still in its infancy. Mimicking the natural evolutionary process of the survival of the fittest, Genetic Algorithms (GAs) have attracted much attention in solving difficult scheduling problems in recent years. Some obstacles exist when using GAs: there is no canonical mechanism to deal with constraints, which are commonly met in most real-world scheduling problems, and small changes to a solution are difficult. To overcome both difficulties, indirect approaches have been presented (in [1] and [2]) for nurse scheduling and driver scheduling, where GAs are used by mapping the solution space, and separate decoding routines then build solutions to the original problem. In our previous indirect GAs, learning is implicit and is restricted to the efficient adjustment of weights for a set of rules that are used to construct schedules. The major limitation of those approaches is that they learn in a non-human way: like most existing construction algorithms, once the best weight combination is found, the rules used in the construction process are fixed at each iteration. However, normally a long sequence of moves is needed to construct a schedule and using fixed rules at each move is thus unreasonable and not coherent with human learning processes. When a human scheduler is working, he normally builds a schedule step by step following a set of rules. After much practice, the scheduler gradually masters the knowledge of which solution parts go well with others. He can identify good parts and is aware of the solution quality even if the scheduling process is not completed yet, thus having the ability to finish a schedule by using flexible, rather than fixed, rules. In this research we intend to design more human-like scheduling algorithms, by using ideas derived from Bayesian Optimization Algorithms (BOA) and Learning Classifier Systems (LCS) to implement explicit learning from past solutions. BOA can be applied to learn to identify good partial solutions and to complete them by building a Bayesian network of the joint distribution of solutions [3]. A Bayesian network is a directed acyclic graph with each node corresponding to one variable, and each variable corresponding to individual rule by which a schedule will be constructed step by step. The conditional probabilities are computed according to an initial set of promising solutions. Subsequently, each new instance for each node is generated by using the corresponding conditional probabilities, until values for all nodes have been generated. Another set of rule strings will be generated in this way, some of which will replace previous strings based on fitness selection. If stopping conditions are not met, the Bayesian network is updated again using the current set of good rule strings. The algorithm thereby tries to explicitly identify and mix promising building blocks. It should be noted that for most scheduling problems the structure of the network model is known and all the variables are fully observed. In this case, the goal of learning is to find the rule values that maximize the likelihood of the training data. Thus learning can amount to 'counting' in the case of multinomial distributions. In the LCS approach, each rule has its strength showing its current usefulness in the system, and this strength is constantly assessed [4]. To implement sophisticated learning based on previous solutions, an improved LCS-based algorithm is designed, which consists of the following three steps. The initialization step is to assign each rule at each stage a constant initial strength. Then rules are selected by using the Roulette Wheel strategy. The next step is to reinforce the strengths of the rules used in the previous solution, keeping the strength of unused rules unchanged. The selection step is to select fitter rules for the next generation. It is envisaged that the LCS part of the algorithm will be used as a hill climber to the BOA algorithm. This is exciting and ambitious research, which might provide the stepping-stone for a new class of scheduling algorithms. Data sets from nurse scheduling and mall problems will be used as test-beds. It is envisaged that once the concept has been proven successful, it will be implemented into general scheduling algorithms. It is also hoped that this research will give some preliminary answers about how to include human-like learning into scheduling algorithms and may therefore be of interest to researchers and practitioners in areas of scheduling and evolutionary computation. References 1. Aickelin, U. and Dowsland, K. (2003) 'Indirect Genetic Algorithm for a Nurse Scheduling Problem', Computer & Operational Research (in print). 2. Li, J. and Kwan, R.S.K. (2003), 'Fuzzy Genetic Algorithm for Driver Scheduling', European Journal of Operational Research 147(2): 334-344. 3. Pelikan, M., Goldberg, D. and Cantu-Paz, E. (1999) 'BOA: The Bayesian Optimization Algorithm', IlliGAL Report No 99003, University of Illinois. 4. Wilson, S. (1994) 'ZCS: A Zeroth-level Classifier System', Evolutionary Computation 2(1), pp 1-18.

An analysis of scheduling algorithms in multiprocessor computing systems /

"UIUCDCS-R-75-724"

Multi-agent system for distributed manufacturing scheduling with genetic algorithms and tabu search

Computerized scheduling methods and computerized scheduling systems according to exemplary embodiments. A computerized scheduling method may be stored in a memory and executed on one or more processors. The method may include defining a main multi-machine scheduling problem as a plurality of single machine scheduling problems; independently solving the plurality of single machine scheduling problems thereby calculating a plurality of near optimal single machine scheduling problem solutions; integrating the plurality of near optimal single machine scheduling problem solutions into a main multi-machine scheduling problem solution; and outputting the main multi-machine scheduling problem solution.

Towards realistic core-failure-resilient scheduling and analysis

Presented at IEEE Real-Time Systems Symposium (RTSS 2015). 1 to 4, Dec, 2015. San Antonio, U.S.A..

Towards realistic core-failure-resilient scheduling and analysis

Presented at IEEE Real-Time Systems Symposium (RTSS 2015). 1 to 4, Dec, 2015. San Antonio, U.S.A..

Real-time Task-Splitting scheduling algorithms framework

Nos dias de hoje, os sistemas de tempo real crescem em importância e complexidade. Mediante a passagem do ambiente uniprocessador para multiprocessador, o trabalho realizado no primeiro não é completamente aplicável no segundo, dado que o nível de complexidade difere, principalmente devido à existência de múltiplos processadores no sistema. Cedo percebeu-se, que a complexidade do problema não cresce linearmente com a adição destes. Na verdade, esta complexidade apresenta-se como uma barreira ao avanço científico nesta área que, para já, se mantém desconhecida, e isto testemunha-se, essencialmente no caso de escalonamento de tarefas. A passagem para este novo ambiente, quer se trate de sistemas de tempo real ou não, promete gerar a oportunidade de realizar trabalho que no primeiro caso nunca seria possível, criando assim, novas garantias de desempenho, menos gastos monetários e menores consumos de energia. Este último fator, apresentou-se desde cedo, como, talvez, a maior barreira de desenvolvimento de novos processadores na área uniprocessador, dado que, à medida que novos eram lançados para o mercado, ao mesmo tempo que ofereciam maior performance, foram levando ao conhecimento de um limite de geração de calor que obrigou ao surgimento da área multiprocessador. No futuro, espera-se que o número de processadores num determinado chip venha a aumentar, e como é óbvio, novas técnicas de exploração das suas inerentes vantagens têm de ser desenvolvidas, e a área relacionada com os algoritmos de escalonamento não é exceção. Ao longo dos anos, diferentes categorias de algoritmos multiprocessador para dar resposta a este problema têm vindo a ser desenvolvidos, destacando-se principalmente estes: globais, particionados e semi-particionados. A perspectiva global, supõe a existência de uma fila global que é acessível por todos os processadores disponíveis. Este fato torna disponível a migração de tarefas, isto é, é possível parar a execução de uma tarefa e resumir a sua execução num processador distinto. Num dado instante, num grupo de tarefas, m, as tarefas de maior prioridade são selecionadas para execução. Este tipo promete limites de utilização altos, a custo elevado de preempções/migrações de tarefas. Em contraste, os algoritmos particionados, colocam as tarefas em partições, e estas, são atribuídas a um dos processadores disponíveis, isto é, para cada processador, é atribuída uma partição. Por essa razão, a migração de tarefas não é possível, acabando por fazer com que o limite de utilização não seja tão alto quando comparado com o caso anterior, mas o número de preempções de tarefas decresce significativamente. O esquema semi-particionado, é uma resposta de caráter hibrido entre os casos anteriores, pois existem tarefas que são particionadas, para serem executadas exclusivamente por um grupo de processadores, e outras que são atribuídas a apenas um processador. Com isto, resulta uma solução que é capaz de distribuir o trabalho a ser realizado de uma forma mais eficiente e balanceada. Infelizmente, para todos estes casos, existe uma discrepância entre a teoria e a prática, pois acaba-se por se assumir conceitos que não são aplicáveis na vida real. Para dar resposta a este problema, é necessário implementar estes algoritmos de escalonamento em sistemas operativos reais e averiguar a sua aplicabilidade, para caso isso não aconteça, as alterações necessárias sejam feitas, quer a nível teórico quer a nível prá

Algorithms and computational tools for metabolic flux analysis

PhD thesis in Biomedical Engineering

Clustering algorithms and shape factor methods to discriminate among small GTPase phenotypes using DIC image analysis.

Naïvement perçu, le processus d’évolution est une succession d’événements de duplication et de mutations graduelles dans le génome qui mènent à des changements dans les fonctions et les interactions du protéome. La famille des hydrolases de guanosine triphosphate (GTPases) similaire à Ras constitue un bon modèle de travail afin de comprendre ce phénomène fondamental, car cette famille de protéines contient un nombre limité d’éléments qui diffèrent en fonctionnalité et en interactions. Globalement, nous désirons comprendre comment les mutations singulières au niveau des GTPases affectent la morphologie des cellules ainsi que leur degré d’impact sur les populations asynchrones. Mon travail de maîtrise vise à classifier de manière significative différents phénotypes de la levure Saccaromyces cerevisiae via l’analyse de plusieurs critères morphologiques de souches exprimant des GTPases mutées et natives. Notre approche à base de microscopie et d’analyses bioinformatique des images DIC (microscopie d’interférence différentielle de contraste) permet de distinguer les phénotypes propres aux cellules natives et aux mutants. L’emploi de cette méthode a permis une détection automatisée et une caractérisation des phénotypes mutants associés à la sur-expression de GTPases constitutivement actives. Les mutants de GTPases constitutivement actifs Cdc42 Q61L, Rho5 Q91H, Ras1 Q68L et Rsr1 G12V ont été analysés avec succès. En effet, l’implémentation de différents algorithmes de partitionnement, permet d’analyser des données qui combinent les mesures morphologiques de population native et mutantes. Nos résultats démontrent que l’algorithme Fuzzy C-Means performe un partitionnement efficace des cellules natives ou mutantes, où les différents types de cellules sont classifiés en fonction de plusieurs facteurs de formes cellulaires obtenus à partir des images DIC. Cette analyse démontre que les mutations Cdc42 Q61L, Rho5 Q91H, Ras1 Q68L et Rsr1 G12V induisent respectivement des phénotypes amorphe, allongé, rond et large qui sont représentés par des vecteurs de facteurs de forme distincts. Ces distinctions sont observées avec différentes proportions (morphologie mutante / morphologie native) dans les populations de mutants. Le développement de nouvelles méthodes automatisées d’analyse morphologique des cellules natives et mutantes s’avère extrêmement utile pour l’étude de la famille des GTPases ainsi que des résidus spécifiques qui dictent leurs fonctions et réseau d’interaction. Nous pouvons maintenant envisager de produire des mutants de GTPases qui inversent leur fonction en ciblant des résidus divergents. La substitution fonctionnelle est ensuite détectée au niveau morphologique grâce à notre nouvelle stratégie quantitative. Ce type d’analyse peut également être transposé à d’autres familles de protéines et contribuer de manière significative au domaine de la biologie évolutive.

Joint Analysis of Multiple Algorithms and Performance Measures

There has been an increasing interest in the development of new methods using Pareto optimality to deal with multi-objective criteria (for example, accuracy and time complexity). Once one has developed an approach to a problem of interest, the problem is then how to compare it with the state of art. In machine learning, algorithms are typically evaluated by comparing their performance on different data sets by means of statistical tests. Standard tests used for this purpose are able to consider jointly neither performance measures nor multiple competitors at once. The aim of this paper is to resolve these issues by developing statistical procedures that are able to account for multiple competing measures at the same time and to compare multiple algorithms altogether. In particular, we develop two tests: a frequentist procedure based on the generalized likelihood-ratio test and a Bayesian procedure based on a multinomial-Dirichlet conjugate model. We further extend them by discovering conditional independences among measures to reduce the number of parameters of such models, as usually the number of studied cases is very reduced in such comparisons. Data from a comparison among general purpose classifiers is used to show a practical application of our tests.

Data Mining for Network Intrusion Detection : A comparison of data mining algorithms and an analysis of relevant features for detecting cyber-attacks

Data mining can be defined as the extraction of implicit, previously un-known, and potentially useful information from data. Numerous re-searchers have been developing security technology and exploring new methods to detect cyber-attacks with the DARPA 1998 dataset for Intrusion Detection and the modified versions of this dataset KDDCup99 and NSL-KDD, but until now no one have examined the performance of the Top 10 data mining algorithms selected by experts in data mining. The compared classification learning algorithms in this thesis are: C4.5, CART, k-NN and Naïve Bayes. The performance of these algorithms are compared with accuracy, error rate and average cost on modified versions of NSL-KDD train and test dataset where the instances are classified into normal and four cyber-attack categories: DoS, Probing, R2L and U2R. Additionally the most important features to detect cyber-attacks in all categories and in each category are evaluated with Weka’s Attribute Evaluator and ranked according to Information Gain. The results show that the classification algorithm with best performance on the dataset is the k-NN algorithm. The most important features to detect cyber-attacks are basic features such as the number of seconds of a network connection, the protocol used for the connection, the network service used, normal or error status of the connection and the number of data bytes sent. The most important features to detect DoS, Probing and R2L attacks are basic features and the least important features are content features. Unlike U2R attacks, where the content features are the most important features to detect attacks.