In this university habilitation, we have set up a decision support platform based on multi-agent systems. From the inputs of our system we configure an architecture in adequacy with the various needs. We have used various cutting-edge technologies, namely artificial intelligence for diagnostic assistance in the context of medical applications, road applications, low-power embedded systems, etc., the Blockchain in the context of securing and sharing data in the context of Industry 4.0, smart cities, smart agriculture. To minimize energy consumption, we have used approximate computing to show the possibility of having efficient approximate systems for artificial intelligence application.