<![CDATA[This study addresses an enhanced version of the Double Row Layout Problem (DRLP) by incorporating two critical constraints: minimum safety distances between machines and geometric limitations on row lengths. A bi-objective mixed-integer non-linear programming (MINLP) model is formulated to simultaneously minimize material handling costs and penalties for violating safety distance requirements. To solve the problem efficiently, a novel metaheuristic called Improved Multi-Objective Variable Neighborhood Search (IMOVNS) is proposed. IMOVNS extends the standard MOVNS by integrating an adaptive archive update strategy and a probabilistic acceptance mechanism inspired by AMOSA, thereby improving both convergence and diversity in Pareto front generation. This study contributes to the layout optimization literature by proposing a tailored MOVNS variant explicitly designed for safety-aware and geometry-constrained DRLP, a challenging problem variant that has received limited attention in prior research. Extensive experiments on 27 DRLP instances show that IMOVNS demonstrates strong performance, significantly outperforming NSGA-II and showing competitive or superior results compared to AMOSA and MOVNS in terms of convergence and solution diversity. Statistical tests further confirm the significant superiority of IMOVNS, particularly over NSGA-II. Additionally, a key managerial insight reveals that layouts with unbalanced row lengths favour safety compliance, while balanced layouts minimize material handling costs. The Pareto-optimal solutions generated by IMOVNS enable decision-makers to select layout configurations that align with specific operational priorities. These findings highlight the practical relevance and robustness of IMOVNS in solving real-world multi-objective facility layout problems under complex spatial and safety constraints.]]>
