Distributed cooperation can be referred to as a behavior, where interacting entities (agents) adjust their states in accordance with information obtained from “nearest” neighbors in a group. Although this idea has roots in biological populations and social systems, being applied to artificial systems it became one of the main directions in modern control theory and optimization. Implementation of distributed cooperation for control and optimization purposes may reduce the complexity of the entire “global” problem: instead of applying a straightforward approach, the problem can be solved through “local” actions. Moreover, in some applications the whole state vector should not be transmitted due to privacy reasons. Our talk is a brief introduction into multi-agent systems, benchmark consensus problem, basic ideas, and results obtained by our scientific group.