OPTIMUS: Optimization Productivity Tool for Intelligent Management of Utilizable Space
Souvik Bhattacharyya, Nisha Singh, Salman Haider, Balaji Nagarajan, Ved Prakash Dwivedi, Nithin Surendran, Karthik Nair · 2026
Read OPTIMUS: Optimization Productivity Tool for Intelligent Management of Utilizable Space by Souvik Bhattacharyya, Nisha Singh, Salman Haider, Balaji Nagarajan, Ved Prakash Dwivedi, Nithin Surendran, Karthik Nair (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-14430.txt.
Introduction
Bhattacharyya and colleagues study department retail space optimisation, allocating bay capacity among planograms under business constraints. They formulate the problem as a linear binary knapsack model. Potential products become items with space needs and value weighted by sales, margin, units and similarity. Dynamic programming gives exact, reproducible decisions, avoiding heuristic variance. A second-stage bay model as a mixed-integer program integrates those decisions. Across ten runs over departments and store clusters, OPTIMUS lifts sales by 11.8 percent and margin by 9.5 percent on average.
Catalog record
| CREATOR | Souvik Bhattacharyya, Nisha Singh, Salman Haider, Balaji Nagarajan, Ved Prakash Dwivedi, Nithin Surendran, Karthik Nair |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25434 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.14430
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
OPTIMUS: Optimization Productivity Tool for Intelligent Management of Utilizable Space Souvik Bhattacharyya, Nisha Singh, Salman Haider, Balaji Nagarajan, Ved Prakash Dwivedi, Nithin Surendran, Karthik Nair Abstract We study department-level retail space optimization, where limited bay capacity must be allocated among planograms (POGs) under business and operational constraints. The problem is formulated as a linear binary knapsack model, with potential SKUs treated as items characterized by space requirements and weighted value contributions from sales, margin, units, and assortment similarity. Dynamic Programming (DP) is employed to obtain exact and reproducible assortment decisions in O ( n c ) O(nc) time, avoiding the variance inherent in heuristic approaches. These decisions are integrated with a second-stage bay optimization model formulated as a mixed-integer program. Evaluated end-to-end across ten optimization runs spanning multiple departments and store clusters, the OPTIMUS framework achieves an average sales lift of 11.8% and an average margin lift of 9.5%. Overall, OPTIMUS provides a scalable, interpretable, and profit-driven solution for enterprise-scale retail space management. † † affiliationnotice: Equal contribution 1 Introduction In the retail industry, physical space is directly tied to revenue generation, making the efficient utilization of available space a critical challenge. Retail floor areas are typically divided into bays ( 3 ) , which represent the fundamental sellable units where products are displayed. These bays are then aggregated into planograms or POGs – visual merchandising blueprints that define product placement and category layouts across stores. Planograms, in turn, roll upward into broader selling departments that drive profitability. It is at the department level that the problem of space optimization becomes most prominent, as it directly affects both store-level performance metrics and strategic decision-making ( 2 ; 4 ) Retailers operate under increasing pressure to maximize profitability within constrained store footprints. As customer demand for product variety grows, the tension between limited physical space and the need for broad assortments intensifies. This makes space allocation not only an operational task but also a strategic decision that influences category management, supplier negotiations, and overall store performance ( 6 ) . Traditional research in this area has proposed a variety of models for space management ( 11 ) , ranging from demand-based space elasticity frameworks ( 4 ) to heuristic algorithms and mixed-integer programming approaches ( 12 ) While these methods offer valuable insights, they often operate at the bay or shelf level, focusing on micro-optimizations of product arrangements. However, such models can overlook the aggregated effects at the department level, where space allocation decisions yield the greatest managerial impact. Moreover, many existing formulations are nonlinear and computationally intensive, posing practical challenges for implementation in large-scale retail environments. From an operations research perspective, the problem aligns closely with combinatorial optimization. The knapsack problem and its variants have emerged as useful abstractions for modeling resource-constrained allocation problems, including those found in retail ( 7 ) . By adapting this framework to the retail context, it is possible to formalize the trade-off between limited shelf space and the maximization of profit in a mathematically tractable way. Despite substantial research efforts, there remain two critical gaps in the literature. First, most existing studies focus on localized shelf or bay-level optimization without adequately addressing department-level decision-making. Second, many models rely on nonlinear demand functions or complex simulation methods that limit their applicability in practice. This creates a disconnect between academic approaches and the needs of retailers, who require scalable, computationally efficient solutions that can be readily integrated into operational workflows. At the selling department level, where profitability and product strategy converge, the challenge is to allocate limited bay space in a way that maximizes returns while respecting category-level and organizational constraints. Given the increasing complexity of modern retail assortments, solving this optimization problem in a tractable and practical manner is essential. This paper addresses the gap by framing department-level retail space optimization as a linearized binary knapsack problem. In this formulation, each product category represents an item with associated profit potential, while the department’s available bay space serves as the knapsack capacity. The binary structure models inclusion or exclusion decisions, and the linearization ensures computational feasibility for large-scale retail applications. The contributions of this study are threefold: Model Development – We introduce a novel application of the binary knapsack problem to department-level space allocation, bridging the gap between bay-level optimization and higher-level managerial decision-making. Computational Tractability – We propose a linearized formulation that is solvable with standard optimization tools, making it accessible and scalable for real-world retail environments. Managerial Insights – We demonstrate how the model can inform practical trade-offs between product variety, planogram design, and profitability, offering actionable guidance for category management and strategic planning. By combining theoretical rigor with practical applicability, this work contributes to both the operations research literature and the retail industry, providing a framework for addressing one of the sector’s most pressing optimization challenges. 2 Methodology To operationalize the proposed framework, we adopt a two-step optimization approach designed to reflect both the strategic and operational layers of retail space management. The first step, planogram optimization, determines which products should be selected for stocking on shelves. This decision is driven by key performance indicators such as historical sales volume and contribution margins, ensuring that the chosen product set maximizes expected profitability while maintaining assortment relevance. By framing this step as a selection problem, we effectively filter the wide product universe down to the most economically viable candidates for display. Building on this foundation, the second step, bay optimization, addresses the allocation of physical shelf space across selling departments. Here, the focus shifts from item-level selection to space distribution, determining how much of the available bay capacity should be dedicated to each department based on the previously chosen assortments. This step balances the competing demands of different departments, ensuring that space is distributed in proportion to profitability potential while respecting organizational constraints. Together, the two steps create a structured pathway: first identifying the what (products to stock), and then optimizing the how much (space to allocate), thereby aligning product assortment strategies with space utilization for maximum overall returns. 2.1 Planogram optimization A POG is defined as a structured set of items that together represent a specific product category or display unit within a retail store. At the most granular level, we have individual items (for example, different brands or models of refrigerators). When a collection of such related items is grouped together, it forms a planogram—for instance, a Refrigerator POG or a Dishwasher POG. Multiple planograms are then aggregated into a selling department, such as the Appliances Department, which may include refrigerator and dishwasher POGs. To determine how much space should be allocated to a given POG, it is essential to estimate the value generated under different space configurations. The POG optimization problem is therefore defined as finding the optimal subset of items to include within a POG for a specified amount of available space ( 5 ) . This optimization is performed for every possible space allocation and for each POG. The resulting outputs then serve as inputs to the subsequent bay optimization stage, which determines the optimal space distribution among POGs within a department. The total display area available for product placement within a planogram is defined as its