Search

In this presentation, it is described the work leading to the proposal of set-membership techniques to the study of sensitivity analysis.

In this presentation, we discuss the use of Gauss Seidel iterations to achieve faster desynchronization in MAC access in its distributed setting.

Over-parameterized models - a complex systems can have simpler models if possible to measure their effect on the state or output. Multi-agent systems - there are local dynamics coupled by a network and identifying nodes with the most impact is cumbersome. Worst-case effect - proposed solution should look at bad trajectories.

Wireless Sensor Networks (WSNs) - a network composed of nodes using a wireless medium in Time Division Multiple Access (TDMA). No centralized infrastructure - implies the need for a decentralized algorithm to perform desynchronization of transmission. Applicable to surveillance - a group of vigilant robots that want to periodically visit sites to be monitored

The problem of deciding which inputs in a model influence the most the state or output is often of practical importance, especially in the cases in which the system can be over-parameterized. In this context, a designer is required to perform sensitivity analyses so as to select which inputs are the most relevant to the problem at hand and remove those with smaller or no impact. In this paper, we tackle this issue by constructing the exact reachable set of a linear system that relates the inputs with the state of that system. By means of projections and solutions of linear optimization programs, we are able to assess which inputs drive the most the state or the output of a linear system. Illustrative examples are presented in order to provide insights on the proposed method.

Submitted by Ana Rebelo (amsr@isep.ipp.pt) on 2020-11-05T15:00:20Z No. of bitstreams: 1 DM_SilvestreCarvalho_2020_MEI.pdf: 7066879 bytes, checksum: 035435bf774ab610d1098ca02e53e008 (MD5) Approved for entry into archive by Ana Rebelo (amsr@isep.ipp.pt) on 2020-11-05T15:00:55Z (GMT) No. of bitstreams: 1 DM_SilvestreCarvalho_2020_MEI.pdf: 7066879 bytes, checksum: 035435bf774ab610d1098ca02e53e008 (MD5) Made available in DSpace on 2020-11-05T15:00:55Z (GMT). No. of bitstreams: 1 DM_SilvestreCarvalho_2020_MEI.pdf: 7066879 bytes, checksum: 035435bf774ab610d1098ca02e53e008 (MD5) Previous issue date: 2020

Presentation at Conference on Decision and Control 2019.

The OPTool package is an implementation of various state-of-the-art iterative optimization algorithms for differentiable cost functions along with algorithms to solve linear equations. Users can use the toolbox to solve optimization problems, although the code was written to researchers that want to compare their proposals with state-of-the-art implementation. New algorithms can be easily added and the software will be updated to have the most comprehensive list of solvers possible. It also comes with implemented functions to return optimal parameters for these algorithms based on a control-theoretical formulation of the algorithms.

The OPTool package is an implementation of various state-of-the-art it- erative optimization algorithms for di erentiable cost functions along with algorithms to solve linear equations. Users can use the toolbox to solve optimization problems, although the code was written to researchers that want to compare their proposals with state-of-the-art implementation. New algorithms can be easily added and the software will be updated to have the most comprehensive list of solvers possible. It also comes with imple- mented functions to return optimal parameters for these algorithms based on a control-theoretical formulation of the algorithms.

Desynchronization for Decentralized Medium Access Control based on Gauss-Seidel Iterations

Fault Detection for Cyber-Physical Systems: Smart Grid case

Traditional PageRank -Compute relative importance of each webpage in the network. Ranking of Researchers - PageRank has also been proposed to be an alternative to citation count. General Ranking - Since PageRank is equivalent to the relative time spent in a node for a random walk on the graph many more applications are possible.

In the context of fault detection and isolation of Linear Parameter-Varying (LPV) systems, a challenging task appears when the dynamics and the available measurements render the model unobservable, which invalidates the use of standard Set-Valued Observers (SVOs). Two results are obtained in this paper, namely: using a left-coprime factorization, one can achieve set-valued estimates with ultimately bounded hyper-volume and convergence dependent on the slowest unobservable mode; and, by rewriting the SVO equations and taking advantage of a coprime factorization, it is possible to have a low-complexity fault detection and isolation method. Performance is assessed through simulation, illustrating, in particular, the detection time for various types of faults.

This work addresses the problem of finite-time convergence of, and the determination of the factors that impact on, the final opinion in a social network for a political party or an association, modeled as a distributed iterative system with graph dynamics chosen to mimic how people interact. It is firstly shown that, in this setting, finite-time convergence is achieved only when nodes form a complete network, and that contacting with agents with distinct opinions reduces to a half the required interconnections. Two novel strategies are presented that enable finite-time convergence, even for the case where each node only contacts the two closest neighbors. It is shown that, in a deterministic setting, the final opinion depends on a so-called connectivity parameter, which influences the relative contribution of each agent's initial belief. In the stochastic case, analogous conclusions are drawn, but in terms of expected values. The proposed strategies and results are relevant also in the context of mobile robot networks where the same assumption of having nodes communicating to their closest neighbors is satisfied. In addition, the results obtained are relevant in terms of saving resources while ensuring finite-time consensus. The performance of the proposed strategies is evaluated through simulation, illustrating, in particular, the key nodes that drive the network, as well as the associated rate of convergence.

This paper addresses the problem of reaching consensus when the position measurements are corrupted by noise and taken at different time instants. Two different scenarios are considered, namely: when the sensor function is performed on a group of nodes and sent to those nodes at the same time; and a second case where the transmission is unicast but neighboring nodes receive the message due to the shared nature of the wireless medium. A solution involving Set-Valued Observers (SVOs) is proposed to maintain and update the set-valued estimates for the positions and how to drive the agents. The algorithm is proved to enforce convergence of the nodes to clusters up to a factor dependent on the uncertainty. The performance of the proposed algorithm is assessed through simulations, illustrating, in particular how the choice for selecting how to divide the area can reduce or increase the number of clusters.

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) A method for the multielemental determination of metals and nonmetals in goji berries (Lycium barbarum L.) by axially and radially viewed inductively coupled plasma-optical emission spectrometry (ICP-OES) is presented. The accuracy of the entire proposed method was confirmed by three certified reference material analyses, and the certified values showed a good agreement at a 95% confidence limit (Student's Rest). Univ Fed Sao Paulo, Sao Paulo, Brazil Univ Fed Sao Paulo, Inst Quim, Sao Paulo, Brazil Univ Fed Sao Paulo, Sao Paulo, Brazil Univ Fed Sao Paulo, Inst Quim, Sao Paulo, Brazil FAPESP: 2012/11517-1 FAPESP: 2012/11998-0 CNPq: 475282/2013-2 CNPq: 475455/2013-4 Web of Science

Compute the average of the initial states in finite-time. Guarantee fault detection and bounds on the maximum possible deviation from an attack. Incorporate the transmissions stochastic information in the fault detection mechanism.

Compute the average of the initial states using asymmetric and randomized communications. Guarantee almost sure convergence, convergence in mean square sense and convergence in expected value

The coexistence between different types of templates has been the choice solution to the information crisis of prebiotic evolution, triggered by the finding that a single RNA-like template cannot carry enough information to code for any useful replicase. In principle, confining $d$ distinct templates of length $L$ in a package or protocell, whose survival depends on the coexistence of the templates it holds in, could resolve this crisis provided that $d$ is made sufficiently large. Here we review the prototypical package model of Niesert et al. 1981 which guarantees the greatest possible region of viability of the protocell population, and show that this model, and hence the entire package approach, does not resolve the information crisis. This is so because to secure survival the total information content of the protocell, $Ld$, must tend to a constant value that depends only on the spontaneous error rate per nucleotide of the template replication mechanism. As a result, an increase of $d$ must be followed by a decrease of $L$ to ensure the protocell viability, so that the net information gain is null., 11 pages, two columns, 11 figures, submitted to J. Theor. Biol