Aimed at the problem that traditional decoupling control method is not ideal for the decoupling effect of nonlinear, multi variable and strong coupling systems, we put forward a control method combining the cuckoo search algorithm (MCS) which use learning evolution as search guide with PID neural network, it eliminate the strong coupling effect between the loops through the PIDNN network training and use the improved cuckoo algorithm to optimize the connection weights. Compared with the reverse BP neural network, the MCS algorithm overcomes the characteristic that is easy to fall into local convergence of PIDNN, and comparing with basic cuckoo algorithm, it has better ability in search precision and convergence speed. By the combination of learning evolving and Gauss distribution and the new search mechanism formed by mixing it with Levy flight search mechanism based on probability of selection, it accelerate the optimization of PIDNN weights. Through the simulation of boiler combustion system of strong coupling fire power plant, it appears that this method has better control characteristics such as good decoupling performance, strong robustness and high control accuracy, which provides an effective way to improve the actual industrial production process of strong coupling control system.
[1]
Amir Hossein Gandomi,et al.
Cuckoo search algorithm: a metaheuristic approach to solve structural optimization problems
,
2011,
Engineering with Computers.
[2]
Wang Wei,et al.
Survey on shuffled frog leaping algorithm
,
2012
.
[3]
Pinar Civicioglu,et al.
A conceptual comparison of the Cuckoo-search, particle swarm optimization, differential evolution and artificial bee colony algorithms
,
2013,
Artificial Intelligence Review.
[4]
Huang Xin,et al.
Cooperative Co-Evolutionary Cuckoo Search Algorithm for Continuous Function Optimization Problems
,
2013
.
[5]
Minghao Yin,et al.
A hybrid cuckoo search via Lévy flights for the permutation flow shop scheduling problem
,
2013
.
[6]
Nie Hu.
Hybrid differential evolution and cuckoo search algorithm for resource-constrained project scheduling
,
2014
.
[7]
Li Xiao-jie.
Inverse System Control Using New Dynamic Neural Networks
,
2012
.