A Method for Decentralised Optimisation in Networks
نویسنده
چکیده
We outline a method for distributed Monte Carlo optimisation of computational problems in networks of agents, such as peer-to-peer networks of computers. The optimisation and messaging procedures are inspired by gossip protocols and epidemic data dissemination, and are decentralised, i.e. no central overseer is required. In the outlined method, each agent follows simple local rules and seeks for better solutions to the optimisation problem by Monte Carlo trials, as well as by querying other agents in its local neighbourhood. With proper network topology, good solutions spread rapidly through the network for further improvement. Furthermore, the system retains its functionality even in realistic settings where agents are randomly switched on and off.
منابع مشابه
Decentralised Resource Allocation and Coordination for 5G Cellular Communication Networks
In order to cope with the ever increasing traffic load that networks will need to support, a new approach for planning cellular networks deployments should be followed. Traditionally, cell association and resource allocation has been based on the received signal power but this approach seems to be inadequate regarding the brewing of heterogeneous networks. In this work, we first implement a net...
متن کاملLayout Optimisation of Decentralised Energy Systems Under Uncertainty
We present a modelling approach to support the layout optimisation of decentralised energy systems composed of photovoltaic (PV) panels and heat pumps with thermal storage capabilities. The approach integrates the simulation-based generation of model input on the basis of publicly available meteorological data and the subsequent optimisation. Selected results concerning the choice of an appropr...
متن کاملDecentralised Supply Chain Formation: A Belief Propagation-based Approach
Decentralised supply chain formation involves determining the set of producers within a network able to supply goods to one or more consumers at the lowest cost. This problem is frequently tackled using auctions and negotiations. In this paper we show how it can be cast as an optimisation of a pairwise cost function. Optimising this class of functions is NP-hard but good approximations to the g...
متن کاملDecentralised conic optimisation of reactive power considering uncertainty of renewable energy sources
This study proposes a decentralised reactive power optimisation of capacitors in distribution system with uncertain renewable energy sources (RES). The optimisation problem is modelled as minimising the active power loss and installed capacitors costs, subject to power flow constraints and other operation conditions. In view of the non-linear power flow equality constraints with uncertain power...
متن کاملBounded decentralised coordination over multiple objectives
We propose the bounded multi-objective max-sum algorithm (B-MOMS), the first decentralised coordination algorithm for multi-objective optimisation problems. B-MOMS extends the max-sum message-passing algorithm for decentralised coordination to compute bounded approximate solutions to multi-objective decentralised constraint optimisation problems (MO-DCOPs). Specifically, we prove the optimality...
متن کامل