A data-driven genetic algorithm optimization and discrete-event simulation validation framework for supermarket layout redesign

This study develops a data-driven two-stage optimization-simulation framework for supermarket layout redesign in space-constrained retail environments. A constrained spatial optimization model is first formulated to minimize weighted customer travel distance using empirically observed inter-department movement data. A Genetic Algorithm generates feasible layout configurations under practical spatial and operational constraints. The optimized layout is then evaluated using a Discrete-Event Simulation model that captures stochastic customer arrivals, movement behavior, queue formation, and checkout service dynamics. The framework is applied to a neighborhood-scale supermarket in southern Vietnam. The optimized layout reduces the weighted distance score by 26.93% and total customer travel distance by 24.75% compared with the existing layout. Simulation results further show a 21.06% reduction in average customer travel time and a 2.05% increase in throughput. However, average waiting time increases by 117.31%, while total time in the system remains almost unchanged, indicating a shift in the operational bottleneck toward checkout services. Methodologically, the study contributes a data-driven decision-support framework linking empirical customer movement data with constrained layout optimization and stochastic simulation-based validation. The findings demonstrate that layout decisions should be evaluated using multiple operational indicators rather than travel distance alone.