Demand uncertainty can lead to excess inventory holdings, capacity creation, emergency deliveries, and stock-outs. The costs of demand may be directly borne by upstream suppliers, but propagate downstream in the form higher prices. To address these problems, we investigate a practical application fixed order commitment contract (FOCC) which manufacturer commits minimum quantity each period receives per unit price discount from supplier for commitment. We model FOCC as Stackelberg game offers...

This paper presents a novel hybrid rolling-horizon strategy to address the dynamic rebalancing problem in large-scale, free-floating bike-sharing systems (FFBSSs). The involves determining routes and quantity of bikes be loaded unloaded at various locations minimize vehicle costs reduce degree imbalance system. proposed consists two stages: preplanning stage that solves preplanned scheme based on historical data, real-time compute operation over rolling horizon. We propose candidate grid...

Deploying deep reinforcement learning (DRL) in real-world inventory management presents challenges, including dynamic environments and uncertain problem parameters, e.g. demand lead time distributions. These challenges highlight a research gap, suggesting need for unifying framework to model solve sequential decision-making under parameter uncertainty. We address this by exploring an underexplored area of DRL management: training generally capable agents (GCAs) zero-shot generalization...

Download This Paper Open PDF in Browser Add to My Library Share: Permalink Using these links will ensure access this page indefinitely Copy URL DOI

Synthetic fingerprint generation has two major advantages. First, it is possible to create arbitrarily large databases for research purposes e.g. of a million or billion fingerprints at virtually no cost and without legal constraints. Secondly, together with the generated images comes additional ground truth information free such as corresponding minutiae template. However, recently been shown that existing methods in literature synthesize unrealistic configurations, usually not visible...

We study inventory control with volume flexibility: A firm can replenish using period‐dependent base capacity at regular sourcing costs and access additional supply a premium. The optimal replenishment policy is characterized by two base‐stock levels but determining their values not trivial, especially for nonstationary correlated demand. propose the Lookahead Peak‐Shaving that anticipates peak shaves orders from future peak‐demand periods to current period, thereby matching Peak shaving...

In many practical settings, a planner must review and optionally adjust recommendations from decision support system. When the system is well-tuned to its task, adjustments can be rare could even degrade system's recommendation. To address these inefficiencies, we develop machine learning (ML) classification model that predicts when will make value-enhancing system-generated of part order quantities. The remaining automated, saving planner's time improving performance by avoiding adverse...

We propose a novel fingerprint descriptor, namely M\"obius moduli, measuring local deviation of orientation fields (OF) fingerprints from conformal fields, and we method to robustly measure them, based on tetraquadrilaterals approximate modulus locally with one due transformation. Conformal arise by the approximation OFs given zero pole models, which are determined singular points rotation. This is very coarse, e.g. for no (arch type), zero-pole model's OF has parallel lines. Quadratic...

We study an inventory model with volume flexibility. A firm may replenish its by using base capacity at regular sourcing cost, after which additional supply is available a premium. The structure of the optimal replenishment policy known to be modified base-stock policy, yet one must resort computational approaches set levels. This non-trivial in non-stationary and correlated environments, thus companies often rely on simplified, suboptimal heuristics. employ robust optimization leverage...

Decision support systems for supply chain planning have been supporting planners over decades to improve their decision making in many different ways. As next-generation are leveraging advanced artificial intelligence (AI) technologies, companies must not only determine what use, but effectively shape how the planner (“the human”) and system machine”) work together. At same time, AI-supported will change job profiles required skill sets of planners. This chapter provides guidance on consider...