Self-organisation provides a suitable paradigm fordeveloping self-managed complex distributed systems, such asgrid computing and sensor networks. In this paper, an integrativeself-organisation mechanism is proposed. Unlike currentrelated studies, which propose only a single principle of selforganisation,this mechanism synthesises the three principlesof self-organisation: cloning/spawning, resource exchange andrelation adaptation. Based on this mechanism, an agent can autonomouslygenerate new agents when it is overloaded, exchangeresources with other agents if necessary, and modify relationswith other agents to achieve a better agent network structure.In this way, agents can adapt to dynamic environments. Theproposed mechanism is evaluated through a comparison withthree other approaches, each of which represents state-of-theartresearch in each of the three self-organisation principles.Experimental results demonstrate that the proposed mechanismoutperforms the three approaches in terms of the profit of individualagents and the entire agent network, the load-balancingamong agents, and the time consumption to finish a simulationrun.