AI Based Quantitative Optimization Models for FMCG Supply Chain Efficiency in High-Demand Markets: A Linear Programming and Mixed-Integer Programming Approach

Abstract

This study addressed a persistent supply chain performance gap in fast moving consumer goods environments where AI forecasting and visibility tools exist, yet efficiency declines under demand surges because replenishment and distribution decisions are not optimized under real capacity and lead time constraints. The purpose was to quantify how AI enabled analytics capability and related planning practices influence supply chain efficiency and to validate the statistical results with prescriptive optimization outcomes. A quantitative, cross sectional, case-based design was used, combining survey evidence from 168 supply chain professionals with cloud and enterprise planning cases modeled under normal and high demand conditions. Key variables included AI analytics capability, forecasting effectiveness, inventory optimization practice, logistics and distribution planning quality, supplier coordination and lead time reliability. The analysis plan applied reliability testing, descriptive statistics, Pearson correlation, and multiple regression, followed by linear programming and mixed integer programming scenario evaluation with feasibility checks and forecast noise sensitivity testing. Reliability was acceptable across constructs (Cronbach alpha 0.79 to 0.88). Descriptive results showed moderately high AI analytics capability (M 3.84, SD 0.62) and forecasting effectiveness (M 3.71, SD 0.66), while overall efficiency remained moderate (M 3.52, SD 0.67). Supply chain efficiency correlated strongly with AI analytics capability (r 0.62, p < 0.001). The regression model explained 54 percent of variance in efficiency (R2 0.54, p < 0.001), with AI analytics capability as the strongest predictor (beta 0.34, p < 0.001), followed by forecasting effectiveness (beta 0.21, p = 0.001), inventory optimization practice (beta 0.17, p = 0.005), and supplier coordination (beta 0.14, p = 0.017); logistics planning was positive but not statistically significant (p = 0.075). Optimization results reinforced these findings: under normal demand, total cost decreased 11.8 percent (1.72M to 1.52M), service level improved from 92.1 to 96.0 percent, and stockouts reduced 18.4 percent; under high demand, cost decreased 9.3 percent (2.05M to 1.86M), service improved from 88.4 to 93.2 percent, and stockouts reduced 15.1 percent, with a 0.8 percent optimality gap and solutions remaining feasible across scenarios. With plus or minus 10 percent forecast noise, cost rose only 2.1 percent and service declined 0.9 points, indicating robustness. The findings imply that enterprises should prioritize analytics and forecasting governance, enforce disciplined inventory policy execution, and embed constraint-based optimization into routine planning to sustain cost and service performance during demand peaks.