Cooperative control in multi-agent systems enable the implementation of a "global" objective through the use of "local" regulatory strategies in networked systems. To account for the coupling of the subsystems or agents within the control, status data are typically exchanged between neighboring agents (see figure). For this purpose, status data is typically exchanged between individual neighboring agents. Calculations are carried out by neighboring agents as shown in the figure below. Often these are sensitive data (for example e.g. in power grids or autonomous vehicle networks), which have to be protected against inspection by third parties such as neighboring agents.

It is obvious that encrypted communication of the data does not provide comprehensive protection per se, as the state data on neighbouring agents is processed unencrypted. Therefore, the encrypted cooperative control goes a step further and encrypts the local control algorithms as well.

Characteristics of the new technology are a redistribution of the calculation of the local control shares. The secure evaluation of these shares with techniques of homomorphic encryption. The implementation of the technology takes place in two stages. Prior to commissioning of the cooperative regime, portions of the local regulations are identified which depend only on conditions of the neighboring agents. These shares dominate all common cooperative arrangements. For instance, shares are then implemented encrypted on the neighboring agent that the results can only be decrypted by the target agents. During runtime of the control, the encrypted control shares are exchanged between the agents instead of the usual status data. Three security guarantees are thus obtained: (1) the communication is encrypted, (2) neighbor states are protected from inspection and (3) neighboring regulatory shares and strategies are not disclosed.