Imperial College London > Talks@ee.imperial > Featured talks > Learning and Adaptation for Self-Organised Resource Allocation in Open Systems

Learning and Adaptation for Self-Organised Resource Allocation in Open Systems

Add to your list(s) Download to your calendar using vCal

If you have a question about this talk, please contact nobody.

Many open systems, for a variety of applications in distributed computing, networking and manufacturing, face a similar problem: how to collective and distribute a common pool of resources in the absence of a centralised authority. One way is for the system components to mutually agree upon a set of conventional rules, and voluntarily regulate and constrain their behaviour in order to comply with the rules. Indeed, this is just the way that people do it, forming what political scientist (and Nobel Laureate) Elinor Ostrom called self-governing institutions. Furthermore, she not only identified the common features of successful (sustainable) common-pool resource management, but she also turned these into institutional design principles. In previous work, these institutional design principles have been expressed as rules of computational logic, which provides an executable specification for algorithmic governance of an open system. Our simulation experiments have shown how such mutable specifications can be self-organised to provide sustainable and inclusive resource allocation in open systems. So far, so good; but a number of open questions remain.

For example, one of the design principles is that rules should be congruent with the prevailing state of the environment. But how does a set of agents come to know that a particular configuration of the rules is indeed “congruent” with the environment, or that one configuration is “better” than another?

This is one area where machine learning can be of significant assistance, and the main focus of this talk is on how learning and adaptive algorithms can be applied in self-organised resource allocation, for example in a framework for distributive justice, social capital for cooperative decision-making (work of Patricio Petruzzi), interactional justice for self-assessment of fairness (work of David Kurka), conflict resolution (work with Jie Jiang and Ada Diaconescu), and other open issues.

This talk is part of the Featured talks series.

Tell a friend about this talk:

This talk is included in these lists:

Note that ex-directory lists are not shown.

 

Changes to Talks@imperial | Privacy and Publicity