Nowadays, the coronavirus pandemic has and is still causing large numbers of deaths and infected people. Although governments all over the world have taken severe measurements to slow down the virus spreading (e.g., travel restrictions, suspending all sportive, social, and economic activities, quarantines, social distancing, etc.), a lot of persons have died and a lot more are still in danger. Indeed, a recently conducted study [1] has reported that 79% of the confirmed infections in China were caused by undocumented patients who had no symptoms. In the same context, in numerous other countries, since coronavirus takes several days before the emergence of symptoms, it has also been reported that the known number of infections is not representative of the real number of infected people (the actual number is expected to be much higher). That is to say, asymptomatic patients are the main factor behind the large quick spreading of coronavirus and are also the major reason that caused governments to lose control over this critical situation. To contribute to remedying this global pandemic, in this article, we propose an IoT ^a investigation system that was specifically designed to spot both undocumented patients and infectious places. The goal is to help the authorities to disinfect high-contamination sites and confine persons even if they have no apparent symptoms. The proposed system also allows determining all persons who had close contact with infected or suspected patients. Consequently, rapid isolation of suspicious cases and more efficient control over any pandemic propagation can be achieved.

Many of self-stabilizing algorithms have been proposed in literature to deal with fault-tolerance in distributed systems. Most existing works have utilized random graphs (Erdos-Renyi networks) to simulate self-stabilizing algorithms. In the present paper, we propose the use of self-stabilizing algorithms on scale-free graphs (Barabasi-Albert networks) which are more representative for real networks. After that, we test these algorithms under evolutionary dynamic graphs. Performance is evaluated using extensive simulations where three well known self-stabilizing algorithms are tested: nodes coloring, minimal dominating set and maximal independent set. 

La conception d’un réseau local sans-fil (WLAN) constitue un problème d’optimisation difficile. Il s’agit de déterminer les positions des points d’accès (PA) à l’intérieur d’une construction urbaine de manière à augmenter le débit et réduire le nombre de PA. L’optimisation de ces deux objectifs contradictoires nécessite l’utilisation d’une approche multi-objectif. Dans ce livre, on propose d’utiliser les algorithmes génétiques hiérarchisés (AGH) multi-objectif comme heuristique pour la résolution d’un tel problème. Les résultats de simulation montrent que l’utilisation de l’approche proposée offre plus d’efficacité qu’un algorithme génétique standard. La vitesse de convergence vers les solutions Pareto optimales devient plus rapide. Ainsi la distribution des solutions optimales sur le front Pareto est plus uniforme. L'algorithme proposé est caractérisé notamment par sa capacité de s’élargir sur les extrémités du front Pareto optimal au cours du processus d’optimisation en couvrant une grande partie du front Pareto à la fin du processus.

This paper describes an algorithm for graph isomorphism problem. A consistency rule is proposed to detect as soon as possible the isomorphism permutation. The algorithm, called CRGI, tries to find an isomorphism between two input graphs through a backtracking exploration that uses a proposed consistency rule to prune the tree-search. This rule is based on changing cases positions of one adjacency matrix to obtain exactly the second adjacency matrix, according to a permutation that must be defined. If such permutation exists, an isomorphism is detected. The proposed rule is able to prune as early as possible unfruitful branches of the tree-search which leads to reduce the practical time complexity. Experimental results comparing CRGI with other popular algorithms show the effectiveness of CRGI especially for random graphs and trees.