Optimizing the Spatial Configuration and Performance of Upper Floors in Shopping Centers Using Generative Design and the TOPSIS Technique: A Case Study of the G2 Floor of Iran Mall

Abstract

This study aimed to generate redesign alternatives based on previously extracted soft design rules, evaluate and rank the alternatives using the TOPSIS multi-criteria decision-making technique, and simulate the performance of the optimal alternative to improve space-syntax and footfall indicators on the G2 floor of Iran Mall. A quantitative, algorithmic, simulation-based design was employed. Six soft design rules derived from the preceding research phase were parameterized within predefined ranges of physical modification, resulting in five distinct spatial alternatives. The alternatives were evaluated using five key performance indicators—footfall, space occupancy rate, user satisfaction, social interactions, and economic value—and three key risk indicators—remaining blind spots, deviation from predicted footfall, and budget deviation. Criterion weights were determined using the Analytic Hierarchy Process, after which the alternatives were ranked using TOPSIS. Sensitivity analysis and expert feedback were subsequently applied to refine the leading alternative and develop the optimized Alternative +3. Spatial, visual, network, and pedestrian-flow analyses were conducted using DepthmapX, Gephi/AGraph, and MassMotion. TOPSIS identified Alternative 3 as the leading solution, with a relative closeness coefficient of 0.766, a distance of 0.234 from the positive ideal solution, and a distance of 0.766 from the negative ideal solution. Sensitivity analysis confirmed the stability of this selection in six scenarios; only reducing the occupancy-rate weight from 0.25 to 0.20 caused Alternative 4 to replace it. Following further optimization, Alternative +3 achieved a mean integration value of 21.60, mean spatial depth of 4.35 steps, choice value of 5,450, visual permeability of 0.55, only five remaining blind spots, and daily footfall of 5,750 users. Space occupancy reached 82%, while the final TOPSIS score increased to 0.790. Integrating generative design, space-syntax analysis, and TOPSIS provides a systematic and evidence-based framework for optimizing large-scale shopping-center layouts. The optimized configuration demonstrates that restructuring circulation loops, strengthening horizontal and vertical connectivity, and enhancing visual permeability can simultaneously reduce spatial isolation, improve movement distribution and accessibility, and enhance the commercial performance of upper retail floors.