The output power of a photovoltaic module array is affected by daylight intensity and temperature and thus exhibits a nonlinear characteristic. Therefore, a maximum power point tracking (MPPT) controller can be used to maintain the output power of such an array at the highest level. Currently, common tracking methods include the voltage feedback, constant voltage, power feedback, perturb and observe (P&O), and incremental conductance (INC) methods [

1,

2,

3,

4,

5]. In particular, the voltage feedback method [

1,

2] requires first measuring the voltage at the maximum power point of an array under a specific temperature. This method is advantageous for simple interpretation but, when the atmospheric conditions change substantially, it cannot track the updated maximum power point. On the basis of the characteristic that the maximum power point is associated with similar voltage under various irradiations, the constant voltage method [

1] tracks the maximum power point and involves easy control and simple calculation. However, this method also cannot track the updated maximum power point after the atmospheric conditions change substantially. The power feedback method [

3] adopts the variation rates of the output power and voltage (

dP/

dV) to determine the maximum power point. This method decreases energy consumption and exhibits high overall efficiency, but the accuracy of the sensory modules involved is undesirable. The P&O method [

4] perturbs a photovoltaic system by increasing or decreasing the voltage during a fixed cycle. When the output power is increased with increasing voltage during a cycle, the subsequent cycle involves further increasing the voltage; otherwise, the voltage is reduced. This method is advantageous for a simple framework and few measurement parameters, but it cannot accurately track the maximum power point. In addition, the solution obtained using the P&O method can oscillate near the maximum point, thus increasing tracking losses. The INC method [

5] compares the static conductance and dynamic conductance of a photovoltaic array to determine the tracking direction. This method enables accurate control and fast responses but involves high computation and expensive controllers. Hence, a simpler fast-converging technique [

6] is proposed to improve the drawbacks of INC method. The aforementioned maximum power point tracking method is inadequate for examining a photovoltaic module array with shaded or malfunctioning modules because the power–voltage (P–V) curve of the array can exhibit multiple peaks [

7]; hence, using the aforementioned convention tracking methods might mistakenly identify a regional maximum point as the global maximum point.

In recent years, various scholars have proposed smart maximum power tracking techniques for photovoltaic module arrays, including the fuzzy control (FC) [

8,

9], genetic algorithm (GA) [

10], neural network (NN) [

11], artificial neural network (ANN) [

12,

13], and ant colony algorithm (ACA) [

14,

15] methods. In particular, the FC, NN, and ANN methods involve complex control processes and computation; hence, they are difficult to execute. The GA and ACA methods [

10,

14,

15] can only be used to examine photovoltaic module arrays with single-peak output characteristic curves or nonshaded modules. The work proposed in [

16] suggests a simple relationship to predict the correct position of the global maximum power point. However, this method can only be used to examine PV module array with two-peak output characteristics. A maximum power point tracking scheme based on the ripple correlation control (RCC) algorithm was proposed in [

17]. However, this method can only be applied for multilevel inverters. Compared with the conventional methods, these methods enable higher success rates of identifying the global maximum power point.

Because the aforementioned MPPT techniques can only be used to examine photovoltaic module arrays with single-peak or two-peak output characteristic curves, this study employed and modified a particle swarm optimization (PSO) method. Under dissimilar shading ratios, the modified PSO method enabled effective identification of the global maximum power point of a photovoltaic module array with double-peak, triple-peak, and quadruple-peak P–V curves.