Table 5 Investigation of heat source optimization methods for substrate board mounting

Shintaro et al. [66]

Genetic algorithm and Lagrange multiplier

Analysis revealed that the used technique was the most effective way to optimize size under various settings and characteristics.

The number of operational machines will be constrained in the intermediate season because of the lesser loads.

Tao et al. [67]

Genetic Algorithm II

The results showed that a fluctuation of roughly 85.53% in the skin friction coefficient was achieved with a variance of 17.19% in the HT coefficient.

There was a very stringent space limitation for a battery pack.

T. K. Hotta, et al. [68]

Artificial neural network (ANN) with a genetic algorithm

According to the findings, the combination of ANN and GA was shown to be quick in achieving the HS’s global ideal location.

Due to time constraints and available features, geometric parameters were not completely explored.

Yunfei et al. [69]

Genetic algorithm

Observations revealed that multi-objective when compared to other algorithms, the best model displayed a great energy economy.

The performance of the heat sink in terms of heat transfer and hydraulics could not be optimized by a single objective.

Qian et al. [70]

ANN

The employed hybrid modeling technique allowed for the containment of additional effects, such as system health state and current into the battery, and it was also easily expanded to more intricate systems.

Nine buried neurons’ experimental data could not be precisely matched.