Multi-agent Healthcare Information System on Hadoop

摘要

The healthcare industry generates a large amount of data by keeping patients' medical history and due to the diversity of clinical medical equipment. In this paper we address problems that exist in the organ transplantation medical field. Some matching algorithms (based on centralized data) were already presented but expanding the system when it comes to Big Data was not considered before. Problems that may occur in the context of Hadoop are related to uneven data distribution and MapReduce processing. The question that arise is whether a centralized graph algorithm can be adapted to MapReduce in order that the results to be equivalent to centralized processing and efficiency to grow through parallel processing. Our solution uses intelligent agents to collect and process medical data. In case of the matching algorithm, the distributed version can approximate a solution obtained by using the centralized application, but it is more effective as a response time in the Big Data context.