Parameter estimation of static solar photovoltaic models using Laplacian Nelder-Mead hunger games search

Opposite Normalized Trust-Region Reflective (ONTRR): A New Algorithm for Parameter Extraction of Single, Double and Triple Diode Solar Cell Models

“…Secondly, the IEF-based objective function is prone to model degeneration. One can see from Table 2 that the TDM parameters satisfy n 1 = n 2 , which can also be found for the LNMHGS [27] algorithm. In addition, the TDM parameters extracted by the HDE [22] algorithm meet n 1 = n 3 , while the DPDE [21] algorithm yields I 03 ≈ 0.…”

“…Although accurate algorithms exist to solve Equation ( 21), they usually involve the computation of all eigenvalues of H and a Newton process for the secular equation (27). To circumvent costly solving Equation (27) in the full-space of dim variables and be beneficial for large-scale problems, an approximate subspace trust-region interior reflective algorithm [32] was proposed to restrict the full-space sub-problem to a 2-D subspace S. Once S is determined, it is trivial to solve the subspace sub-problem Equation (28) since it is just two-dimensional.…”

“…The optimal results extracted from the IEF-based objective function by the proposed ONTRR are summarized in Table 2. To ensure that the reader can reproduce the RMSE values, the parameter values in Table 2 are correct to 8 decimal places and compared to the results of some state-of-the-art algorithms, such as EJBO [5], GSK [6], GNDO [7], CLRao−1 [8], IGSK [9], IMPA [10], EOTLBO [11], ELBA [12], QRMSCS [13], MIGTO [14], CPMPSO [15], GOFANM [16], WLCSODGM [17], ISCE [18], DOLADE [19], Rcr-IJADE [20], DPDE [21], HDE [22], EJADE [23], ISMA [24], DE/WOA [25], QNMPO [26], LNMHGS [27], EHA-NMS [28] and ABC-TRR [29]. By crosschecking Table 2 and the results in , one can find that there is a less than 0.01 percent differences between them, even though they include different numbers of decimal places.…”

“…2 of 21 swarm [17] and evolution [18][19][20][21][22][23], usually require significant computational efforts to converge and depend strongly upon the proper tuning of control parameters, such as population size, the maximum number of function evaluations (MaxNFEs) or iterations, searching strategies and so on. Hybrid algorithms [24][25][26][27][28][29] combine the strengths of two or more different methods in an attempt to achieve better results. For example, the ABC-TRR [29] algorithm combined the global exploration capability of the artificial bee colony (ABC) algorithm and the local exploitation ability of trust-region reflective (TRR) search to extract the parameters of SDM and DDM.…”

Parameter identification of solar photovoltaic models by multi strategy sine–cosine algorithm

“…In recent years, with the rapid development of society and economy, the human's demand for energy has been increasing day by day. 1 Coal-and oil-based fossil energy will produce a large number of pollutants such as carbon dioxide, sulfur oxides, and nitrogen oxides during the use process, resulting in inestimable environmental losses and inevitably causing frequent occurrences of extreme climates. Therefore, it is extremely urgent to find a clean and pollution-free alternative energy source.…”