Improved Teaching Learning Based Optimization and Its Application in Parameter Estimation of Solar Cell Models

Weak global exploration capability is one of the primary drawbacks in teaching learning based optimization (TLBO). To enhance the search capability of TLBO, an improved TLBO (ITLBO) is introduced in this study. In ITLBO, a uniform random number is replaced by a normal random number, and a weighted average position of the current population is chosen as the other teacher. The performance of ITLBO is compared with that of five meta-heuristic algorithms on a well-known test suite. Results demonstrate that the average performance of ITLBO is superior to that of the compared algorithms. Finally, ITLBO is employed to estimate parameters of two solar cell models. Experiments verify that ITLBO can provide competitive results.