<![CDATA[Reservoir parameter optimisation is a critical aspect of enhancing the hydrocarbon recovery factor and requires systematic, integrated approaches. This research develops an integrated framework combining numerical reservoir modelling with multivariate regression analysis to optimise key reservoir parameters. The research methodology employs quantitative techniques, including three-dimensional numerical simulation, implementation of a multi-objective optimisation algorithm, and development of machine learning models for recovery factor prediction. Research data encompasses reservoir petrophysical parameters, including effective porosity ranging from 12.5% to 28.7%, horizontal permeability from 15 mD to 450 mD, and initial oil saturation from 52.3% to 84.6%. Optimisation analysis using three algorithms demonstrates that the Multi-Objective Grey Wolf Optimiser achieves superior performance, with an optimal recovery factor of 46.3% using only eight parameters. The third-order polynomial regression model yields a coefficient of determination of 0.89 in predicting nonlinear relationships between reservoir parameters and recovery factor. An artificial neural network implementation achieves 94.2% training and 89.3% test prediction accuracy, with a mean absolute error of 2.1%. Development scenario simulation indicates the five-spot injection pattern configuration produces the highest recovery factor, 48.2% with a present value of 187.3 million USD. The developed integrated framework demonstrates the ability to handle reservoir heterogeneity, with validation showing deviations of less than 7% relative to field data.Keywords: numerical modeling, parameter optimization, recovery factor]]>
