Abstract: This article presents an approach for conducting the decision-making process within a company, assisted by intelligent agents. The interest here is centered on the management functions of human resources. The multi agent system has been put in place, in the validation phase for a big brewing enterprise, and is made up of expert agents for each process of HR (Human Resource) in the enterprise. These agents analyze the indicators and produce reports in case there is a problem. This makes it possible for managers to concentrate on the strategic roles on the one hand, leaving part of the follow-up to the agents, and on the other hand to capitalize the knowledge, experiments and expertise within the company.

Abstract: The intelligent water drops (IWD) algorithm was introduced in 2007 by Hamed Shah Hosseini, and has been successfully implemented to solve many types of optimization problems, especially traveling salesman problem (TSP). The algorithm works by imitating the behavior of the natural river water drops. Principally, the algorithm finds solution of the problem by constructing it iteratively in certain number of stages. In this paper, we adjusted (or slightly modified) the IWD algorithm in order to be properly implemented in the flow shop scheduling problem. The testing was conducted on the Taillard (1993) benchmark, and the results were compared with the ones obtained by MOACSA (multi objective ant colony system algorithm). Based on the comparison especially on the makespan values between both algorithms, it is seen that the adjusted IWD algorithm gave better solution than MOACSA on solving flow shop scheduling problem with the objective to optimize the makespan.

Abstract: Peer-to-peer (p2p) systems is a good and cheep alternative for traditional server-client system to providing video streaming for huge number of users; the topology is the one of most important part of these systems. Tree topology is widely used to be the backbone of p2p live video systems in networks’ research or in local networks, mostly in this topology every peer (except leafs) has same number of children peers. This study proposed an innovative algorithm to determine the maximum capacity of children peers based on the farther peer bandwidth and make Max-1 topology applicable in heterogeneous global networks, Max-1 shows 12% better performance than ideal tree topology (guided by human).