Abstract
Large-scale AI data center portfolios procure identical SKUs across geographically heterogeneous campuses, yet finance and operations require a single system-level 'world price' per SKU for budgeting and planning. A common practice is deployment-weighted blending of campus prices, which preserves total cost but can trigger Simpson-type aggregation failures: heterogeneous location mixes can reverse SKU rankings and distort decision signals. I formalize cost-preserving blended pricing under location heterogeneity and propose two practical operators that reconcile accounting identity with ranking robustness and production implementability. A two-way fixed-effects operator separates global SKU effects from campus effects and restores exact cost preservation via scalar normalization, providing interpretable decomposition and smoothing under mild missingness. A convex common-weight operator computes a single set of campus weights under accounting constraints to enforce a location-robust benchmark and prevent dominance reversals; I also provide feasibility diagnostics and a slack-based fallback for extreme mix conditions. Simulations and an AI data center OPEX illustration show substantial reductions in ranking violations relative to naive blending while maintaining cost accuracy, with scalable distributed implementation.

Why we are recommending this paper?
Due to your Interest in: AI for Pricing Optimization

This paper directly addresses pricing strategies within complex supply chains, aligning with your interest in AI for Supply Chain Optimization and Pricing. The focus on blended pricing for data centers, a key area of demand and cost management, is highly relevant.

Abstract
Standard jump-diffusion models assume independence between jumps and diffusion components. We develop a multi-type jump-diffusion model where jump occurrence and magnitude depend on contemporaneous diffusion movements. Unlike previous one-sided models that create arbitrage opportunities, our framework includes upward and downward jumps triggered by both large upward and large downward diffusion increments. We derive the explicit no-arbitrage condition linking the physical drift to model parameters and market risk premia by constructing an Equivalent Martingale Measure using Girsanov's theorem and a normalized Esscher transform. This condition provides a rigorous foundation for arbitrage-free pricing in models with diffusion-dependent jumps.

Why we are recommending this paper?
Due to your Interest in: Pricing

This research explores pricing models, specifically arbitrage-free pricing, which is central to your interest in AI for Pricing Optimization and Pricing. The use of diffusion-dependent jumps offers a sophisticated approach to pricing challenges.

AI Insights

• The LNN+XGBoost hybrid model effectively captures order volatility through the Standard Deviation of Orders over 5 Days variable, enabling strategic adjustments to ordering policies that mitigate upstream demand amplification, a hallmark of the bullwhip effect. [3]
• Bullwhip Effect: A phenomenon where small changes in demand at one level of the supply chain are amplified as they move up the supply chain, leading to large fluctuations in inventory levels and orders. [3]
• The LNN+XGBoost hybrid model outperformed other models in terms of cumulative profits, with a ranking of 1. [3]
• This study compares the performance of different machine learning models for supply chain optimization, including LNN+XGBoost, XGBoost, LSTM, Transformer, and DQN. [3]
• The results show that LNN+XGBoost outperformed other models in terms of cumulative profits, with a ranking of 1. [3]
• This study uses machine learning models to help companies like yours optimize their supply chains and make better decisions about inventory levels. [3]
• The model's ability to capture order volatility and mitigate overstocking contributed to its superior performance. [2]

Abstract
Supply chain management (SCM) faces significant challenges like demand fluctuations and the bullwhip effect. Traditional methods and even state-of-the-art LLMs struggle with benchmarks like the Vending Machine Test, failing to handle SCM's complex continuous time-series data. While ML approaches like LSTM and XGBoost offer solutions, they are often limited by computational inefficiency. Liquid Neural Networks (LNN), known for their adaptability and efficiency in robotics, remain untapped in SCM. This study proposes a hybrid LNN+XGBoost model for multi-tier supply chains. By combining LNN's dynamic feature extraction with XGBoost's global optimization, the model aims to minimize the bullwhip effect and increase profitability. This innovative approach addresses the need for efficiency and adaptability, filling a critical gap in intelligent SCM.

Why we are recommending this paper?
Due to your Interest in: AI for Supply Chain Optimization

This paper tackles supply chain optimization using advanced ML techniques, directly addressing your interest in AI for Supply Chain Optimization and Supply Chain management. The focus on continuous time-series data is particularly relevant.

Abstract
Online Resource Allocation addresses the problem of efficiently allocating limited resources to buyers with incomplete knowledge of future requests. In our setting, buyers arrive sequentially demanding a set of items, each with a value drawn from a known distribution. We study environments where buyers' valuations exhibit complementarities. In such settings, standard item-pricing mechanisms fail to leverage item multiplicities, while existing static bundle-pricing mechanisms rely on problem-specific arguments that do not generalize. We develop a unified technique for online resource allocation with complementarities for three domains: (i) single-minded combinatorial auctions with maximum bundle size $d$, (ii) general single-minded combinatorial auctions, and (iii) a graph-based routing model in which buyers request to route a unit of flow from a source node $s$ to a target node $t$ in a capacitated graph. Our approach yields static and anonymous bundle-pricing mechanisms whose performance improves exponentially with item capacities. For the $d$-single-minded setting with minimum item capacity $B$, we obtain an $O(d^{1/B})$-competitive mechanism, recovering the known $O(d)$ bound for unit capacities ($B=1$) and achieving exponentially better guarantees as capacities grow. For general single-minded combinatorial auctions and the graph-routing model, we obtain $O(m^{1/(B+1)})$-competitive mechanisms, where $m$ is the number of items. We complement these results with information-theoretic lower bounds. We show that no online algorithm can achieve a competitive ratio better than $Ω((m/\ln m)^{1/(B+2)})$ in the general single-minded setting and $Ω((d/\ln d)^{1/(B+1)})$ in the $d$-single-minded setting. In doing so, we reveal a deep connection to the extremal combinatorics problem of determining the maximum number of qualitatively independent partitions of a ground set.

Why we are recommending this paper?
Due to your Interest in: Pricing

This paper investigates resource allocation problems, a core component of supply chain design and optimization. The use of bundle pricing aligns with your interest in pricing strategies within a supply chain context.

AI Insights

• The data generation is performed using an improved FT-based algorithm, termed SOA, developed as an enhancement of CMA. [3]
• The proposed ML-based framework further enhances flexibility and robustness. [3]
• Three ML algorithms including NN, RF and GBDT are trained on the SOA-generated dataset. [3]
• Comparative analyses indicate that all three models achieve high accuracy. [3]
• NN achieves the highest accuracy and the greatest improvement in execution time. [3]
• Although GBDT exhibits slightly lower accuracy and a longer execution time than the NN, it strikes an effective balance among precision, interpretability, and computational efficiency. [3]
• The selection between GBDT and NN should be guided by the nature of the task and the computational capabilities of the available hardware. [3]
• The paper presents an efficient ML-based framework for multiple option pricing tasks where the options are path-independent and the underlying stock prices follow exponential Lévy processes. [2]
• Although an alternative FFT-based framework provides computational acceleration through the FFT and represents a natural extension of SOA (or CMA), it suffers from numerical instability for deep out-of-the-money options and imposes strict requirements on input option attributes. [1]
• SOA leverages the theoretical relationship between the smoothness of a function and the tail decay rate of its Fourier Transform (FT), introducing a smooth offset term that replaces the original offset in CMA. [0]

Abstract
The increasing need for rapid recalibration of option pricing models in dynamic markets places stringent computational demands on data generation and valuation algorithms. In this work, we propose a hybrid algorithmic framework that integrates the smooth offset algorithm (SOA) with supervised machine learning models for the fast pricing of multiple path-independent options under exponential Lévy dynamics. Building upon the SOA-generated dataset, we train neural networks, random forests, and gradient boosted decision trees to construct surrogate pricing operators. Extensive numerical experiments demonstrate that, once trained, these surrogates achieve order-of-magnitude acceleration over direct SOA evaluation. Importantly, the proposed framework overcomes key numerical limitations inherent to fast Fourier transform-based methods, including the consistency of input data and the instability in deep out-of-the-money option pricing.

Why we are recommending this paper?
Due to your Interest in: AI for Pricing Optimization

This paper focuses on option pricing, a critical area within financial modeling and potentially relevant to your interest in AI for Pricing Optimization. The integration of Fourier transforms for efficient recalibration is a valuable technique.