Towards incremental social learning in optimization and multiagent systems