**4. Data analytics models**

A novel integrated model was proposed to improve haul truck energy usage's three most significant and critical effective characteristics. Payload (P), truck speed (S), and total resistance (R) are the three parameters (T.R.). However, the relationship between energy usage and these characteristics on an actual mining site is complicated. Therefore, to predict and reduce haul truck fuel consumption in surface mines, we apply two AI technologies to develop an advanced data analytic model (**Figure 3**).

In the first step, an artificial neural network (ANN) model was developed to create a Fuel Consumption Index (FCIndex) as a function of P, S, and T.R. This index shows how many liters of diesel fuel are consumed to haul 1 ton of mined material in 1 h. In this model, the main parameters used to control the algorithm were R2 and MSE. After the first step, the optimum values of P, S, and T.R. will be determined using a novel multi-objective GA model. These improved parameters can be utilized to boost haul truck energy efficiency.

The proposed model's methods are all based on actual data obtained from surface mines. Below are the results of utilizing the developed model for two genuine major surface mines in Australia and Iran. On the other hand, the finished methods can be expanded for various mines by substituting the data.
