Binary neighbourhood field optimisation for unit commitment problems

Inspired by the neighbourhood cooperation, a new discrete optimisation algorithm is proposed. The so-called binary neighbourhood field optimisation (BNFO), utilises the attractive field of the superior neighbour and the repulsive field of the inferior neighbour. As a kind of local search, BNFO is able to deliver promising results efficiently within acceptable computational time. BNFO is applied to solve the unit commitment problem (UCP), whose objective is to minimise the operation cost of the generation units over the scheduling horizon. After numerical tests on several benchmark UCP cases, the obtained costs are less expensive compared with conventional Lagrangian relaxation, genetic algorithm, evolutionary programming, particle swarm optimisation and differential evolutionary. BNFO can converge to promising results with less computation times, especially for the large-scale UCPs.