In this paper, a parameter identification method of photovoltaic cell model based on improved lion swarm optimization is presented. Lion swarm optimization is a novel intelligent algorithm proposed in recent years, but it has problems such as local optimum and slow convergence. To overcome such limitations, we can combine the tent chaotic map, adaptive parameter and chaotic search strategy to further improve the search ability of the algorithm and avoid trapping in local optimum. The simulation of standard test function shows that the performance of improved lion swarm algorithm is superior to the other six algorithms. Then the algorithm is applied to the parameter identification of photovoltaic cells under two kinds of models and different irradiance, the simulation results verify the superiority and effectiveness of the improved lion swarm optimization in the application of photovoltaic cell parameter identification.

The state of charge (SOC) of lithium batteries is an important parameter of battery management systems. We aim at the problem that the noise variance is fixed during the estimation of the battery state by the unscented Kalman filter (UKF), which leads to low estimation accuracy. Lithium battery SOC estimation based on the UKF and whale optimization algorithm (WOA) is proposed. The first WOA is used to identify the parameters of the battery model. WOA–UKF is used to estimate the SOC of the battery, in which the observed noise variance and process noise variance of the UKF are updated through the second WOA, thereby the estimation accuracy is improved. The experimental results verify the effectiveness of the improved method.

No abstract

<p style='text-indent:20px;'>This paper proposes a multi-objective lion swarm optimization based on multi-agent (MOMALSO) for solving the increasingly complex multi-objective optimization problem in engineering practice. First, the Multi-agent system is introduced into the lion swarm optimization (LSO) algorithm. The optimization mechanism of LSO and the information exchange between the agents are integrated to enhance the local search and global search ability of the algorithm, and the self-learning operation can accelerate the approximate Pareto front obtained by the algorithm near to real front. Besides, the external archive is introduced for extending the LSO into a multi-objective algorithm. Finally, the simulations compared with other three algorithms are performed, and the results show that MOMALSO has significant advantages in both convergence and coverage, which verifies the superiority and effectiveness of the algorithm in multi-objective optimization.</p>

No abstract