Consumer electronic devices have become a part of our everyday lives. A statistically non-negligible percentage of electric loads on the power network is represented by these devices. Consumer electronics notoriously consume DC power, at several voltage levels ranging anywhere from 1.8 to 24 V, in the absence of unified standards. As a consequence, each load is served from the AC network through several conversion stages, which is not the most efficient solution. The problem of efficiency exists not only on a global level, but locally, too. On a local level efficiency can be improved by connecting several converters in parallel and setting their operating point as close to the maximum efficiency point as possible. This can be implemented by means of multi-agent based load sharing algorithms . These intelligent agents (IA) are negotiating at the load distribution between converters and aim at achieving maximum overall efficiency. Other optima can also be targeted at the same time. The paper presents a model based on DC/DC buck converters and its experimental verification.