محمد شاهسون فولادي

Today, the influence of social networks and people and groups active in them is a subject that has always been of interest. The influence of people on each other leads to the dissemination of information in the space of networks. In today's structure of networks, people always belong to different groups due to their interests and type of activities and commonalities with other people. For this reason, eva‎luating the importance of groups and their members is one of the important issues in social networks. Also, the discussion of viral marketing is one of the important and significant issues in small and large institutions and businesses, which is done with the help of social networks. Finding a group of people who can influence the largest part of the network and simultaneously maximize the profit of advertising has been investigated under the title of profit maximization problem. In this research, an approach to find a set of profitable network nodes is presented without the need for prior knowledge; in such a way that they can infiltrate and activate the maximum number of groups in the network and at the same time the cost of advertising does not increase. First, we try to reach certain subgraphs of the main graph with the help of clustering. Then, we identify the profitable nodes of the network by presenting a special evolutionary algorithm called cumulative salps algorithm. So that even if the nodes have different benefit and cost compared to each other, the profit maximization algorithm still has an acceptable performance. In the end, we have compared the proposed algorithm with other algorithms in terms of output quality and execution time.

حداكثرسازي سود, , شبكههاي اجتماعي , شناسايي گروه , بازاريابي ويروسي

Profit Maximization , social networks , community detection

Mohammad Shahsavan Fouladi

Dr. Babak Amiri