*4.1. Data Analysis*

In the third part, the final results of many methods are relatively convergent. For SDM, the RMSE is mainly distributed around 9.8206 × <sup>10</sup>−<sup>4</sup> and the rest is concentrated around 7.7301 × <sup>10</sup>−4. The DDM's primary distribution is around 9.8248 × <sup>10</sup>−4, with a secondary allocation of 7.42 × <sup>10</sup>−<sup>4</sup> to 7.1823 × <sup>10</sup><sup>−</sup>4. For the TDM, the main distribution is between 9.8331 × <sup>10</sup>−<sup>4</sup> and 9.8033 × <sup>10</sup><sup>−</sup>4, with higher precision than the main distribution interval, being 8.2511 × <sup>10</sup>−<sup>4</sup> for TGA and 6.5424 × <sup>10</sup>−<sup>9</sup> for HPSODOX. For Photowatt-PWP201, the main distribution is around 2.4251 × <sup>10</sup>−3, the secondary distribution is around 2.0399 × <sup>10</sup>−3, and the best-performing TSA reaches 5.06 × <sup>10</sup>−4. STM6-40/36 is mainly distributed at 1.7298 × <sup>10</sup><sup>−</sup>3. STP6-120/36 is primarily distributed at 1.6601 × <sup>10</sup>−<sup>2</sup> nearby; the best-performing FDB-TLABC achieved 1.4251 × <sup>10</sup>−2. However, the different approaches rarely use the same cases and evaluation indicators, and the results may differ between models. Therefore, some well-performed variants of metaheuristics that used the RMSE indicators are selected for further comparison in Table 20, i.e., ABC-TRR, RLDE, OLBWOA, CSOOJAYA, DEDIWPSO, EOTLBO, IWOA, TLBOBSA, IGSK, HSOA, and SOS.

**Table 20.** Various methods' RMSE results.


The variants of metaheuristics that used the SIAE indicators are selected for further comparison in Figure 5, i.e., SOS, HSOA, GSK, TLBOBSA, DE/WOA, ISNMWOA, MCSWOA, IWOA, DMTLBO, PSO-ST, GCPSO, MLJAYA, I-GWO, HDE, DPDE, QILDE, ABC-TRR, and TLABC. Moreover, these methods were generally tested in the SDM, DDM, and Photowatt-PWP201 module. Here, the module only means the Photowatt-PWP201.


**Figure 5.** Various methods' Friedman Test.
