How to compare 44 different algorithms to improve energy flow

Links table

Abstract and 1. Introduction

2. Evaluation methods

3. Review the current experiments

4. Circular and application assessments and 4.1. Prediction and response circulation

4.2. Applications for multiple -written situations and 4.3. Zero predict the ability of the application

4.4. Continuous prediction and 4.5. Normalization

5. numerical assessments and 5.1. Experience settings

5.2. Overview of the evaluation

5.3. Failure evaluation

5.4. Accuracy

5.5. Efficiency evaluation

6. Open questions

7. Conclusion

Approach a and references

2. Evaluation methods

Table 1 is prepared 44 methods that have been evaluated, in detail for each of the interviews, the training algorithm used, and any support techniques used. The following points require attention.

First, for the linear programming methods, we also evaluate these methods without their main restrictions, for example, associated restrictions, conjugation restrictions or structure restrictions, in order to verify the added value to integrate these restrictions. It is important to realize that, even without these main restrictions, the resulting programming models remain different, attributed to the various support technologies that combine them.

Second, the first part of this educational program [6] It reveals the standard nature of DPFL studies, highlighting its flexibility in collecting various technologies to form new methodologies. In compliance with this model, we offer many previously unaccredited ways in the DPFL field, and we include them in the following comparative analysis. These methods include small squares with false inverse squares that are increased squares that are increased by the main component analysis, and the smaller squares with Pseudoinves[1]. It is important to make it clear that the goal of merging these methods is not to say “modernity/superiority” over all the technologies present. Instead, we intend to show the ease that one can deviate from traditional paths to create distinctive methodologies. It is worth noting that some of these methods that were presented showed a satisfactory performance and classifications in subsequent assessments. This result, especially given the uncomplicated nature of these methods, indicates a high potential for more developments in DPFL research.

Finally, our evaluation is also surrounded by a selection of physics -based energy flow methods (PPFL), such as the classic DC model, the power distribution factor model, and a first -class Taylor model (derivative of nodal energy injection equations in polar coordinates), and the power flow model Separate linear [7]. Note that these PPFL styles are widely recognized and employed in both research and academic industry.

---

[1] In adapting the micro -squares to integrate PseudoinVerse, the traditional reflection process used in the normal small squares method with the reverse -perose is replaced. This modification is designed to enhance the elasticity of the method of multiple font issues.

Likewise, for the method of microfinances enhanced with the opposite of false, the initial repetition of the well -known well -known method of the use of small squares is adjusted with the opposite of false instead of the regular small squares. This amendment also aims to enhance the ability of the method to manage multiple sin.

In the case of combined partial small squares, the approach includes replacing the component of the regular small squares in the collection -based minor squares (as discussed in the first part [6]) With regular partial small squares. This change seeks to absorb non -linear properties inherent in the frequency power flows.

For details about the small squares with the main component analysis, the reader is referred to the Appendix A.