Multi-agent signal filtering for electrical energy demand management

Consumers participating in electrical energy Demand Response (DR) programs may be exposed to energy-use related decisions at instants of time which are generally hard to predict. This is especially cumbersome to residential consumers who are less capable of investing in special equipment, or devoting significant time to analyze information and take decisions. To ease residential consumer participation, a multi-agent system proposed in this paper filters DR dispatch signals based on participants´ periodic preferences. Thus dispatches appear to consumers in a possibly periodic manner, i.e. at more predictable instants of time from consumer point of view. Additionally, the probability of execution may increase as these instants are defined by consumers themselves. DR needs not matching any participant preferences remain to be covered by other, traditional means. When preference windows are expressed as periodic, DR dispatches appear to these consumers only in these windows, as if dispatches only happen periodically.