Behaving responsible in multi-agent worlds

It has been proposed that a good way of allocating tasks to agents is by ascribing them obligations, i.e., if we want agent i achieves φ, we can stipulate that ‘it is obligatory for i that φ’. Here, we argue that this method is not adequate to guide agent’s decisions. Then, using a multi-agent extension of propositional dynamic logic, with operators expressing agents’ knowledge and abilities, we show that when agents’ decisions are guided by responsibilities, as we define here, a successful performance is more likely to be obtained.