resource allocation

This work addresses the challenge of utility loss under strict resource constraints in machine learning–based decision-making, where conventional fairness criteria can incur unbounded utility degradation and exacerbate inter-group resource disparities. To mitigate this, the authors propose an enhanced proportional fairness criterion alongside a novel variant of equality of opportunity, both of which guarantee bounded fairness-induced utility costs. Building upon algorithmic fairness theory and resource allocation models, they formulate a constrained optimization framework that systematically quantifies utility loss across different fairness definitions. Theoretical analysis and empirical evaluations demonstrate that the proposed approaches significantly improve robustness while preserving fairness, thereby achieving an effective trade-off between accuracy and fairness under resource limitations.

Financial institutions face capacity planning and job scheduling challenges in hybrid cloud and on-premise grid environments, where both resource requirements and execution durations exhibit dual uncertainty. Method: This paper proposes a co-optimization framework that jointly minimizes resource provisioning while maximizing service quality—specifically, on-time completion rate. Innovatively, it is the first to jointly model resource and duration uncertainty within capacity planning, employing a constraint programming framework based on paired sampling that integrates deterministic estimation with stochastic sampling for efficient approximate optimization. Contribution/Results: Experiments demonstrate that the method significantly reduces peak resource demand compared to manual scheduling, while maintaining a high on-time completion rate—validating its effectiveness in balancing these conflicting objectives under uncertainty.

This paper studies dynamic pricing and demand learning under resource constraints, unknown demand distributions, and a hard budget on the number of price changes. A decision-maker must sequentially allocate limited inventory to maximize revenue while respecting both inventory and switching budgets. We establish matching upper and lower regret bounds, revealing that the optimal regret exhibits piecewise-constant dependence on the switching budget—thereby characterizing the fundamental statistical impact of joint resource and switching constraints. We propose a robust algorithm integrating online learning, confidence interval analysis, and adaptive switching control. Theoretically, we prove that regret decays nonlinearly with the switching budget. Empirically, the algorithm achieves near-optimal revenue even when the number of price changes is reduced by 90%, significantly enhancing deployability in industrial settings.

To address resource supply-demand imbalances—manifesting as shortages and surpluses—across globally distributed heterogeneous computing clusters, this paper proposes a resource rationing mechanism grounded in real-world market economics. Methodologically, it introduces a periodic simulated-clock auction framework integrating utilization-driven reserve-price setting, long-term resource quota modeling, and supply-demand equilibrium pricing, enabling dynamic price signals to guide users’ autonomous job placement decisions. Its key contribution lies in being the first to systematically embed microeconomic market mechanisms into large-scale distributed resource allocation, replacing static quota or immediate-scheduling paradigms. Evaluated on the Google experimental market, the mechanism significantly incentivizes user migration toward underutilized clusters: resource utilization variance decreases by 32%, and shortage rate drops by 41%. These results empirically validate that price-based incentives can effectively drive system-level behavioral optimization and achieve global resource equilibrium.

This paper addresses the challenge of ensuring machine learning fairness under resource constraints, where conventional fairness algorithms—ignoring such constraints—often fail in deployment. We propose the first “resource-constrained fairness” framework, formally defining and quantifying the resource cost required to achieve fairness. We reveal that resource availability fundamentally governs fairness cost and establish a unified evaluation and optimization paradigm jointly incorporating resource budgets and group fairness. Methodologically, we model resource consumption via classifier threshold tuning, and jointly optimize fairness metrics (e.g., equalized odds) subject to explicit resource constraints. Our approach integrates theoretical analysis—including derivation of tight theoretical bounds on achievable fairness under adjustable thresholds—with empirical validation. Results demonstrate that resource scarcity substantially increases fairness cost; the derived theoretical boundaries provide actionable design guidelines for fair model deployment in real-world, resource-limited settings.

This study addresses the joint optimization of personnel assignment and performance evaluation rules in networked organizations where employee output depends on both individual effort and exogenous, uncontrollable advantages. The authors develop a dual-network framework: a competition network that shapes effort decisions and a spillover network that propagates advantage effects. By integrating game-theoretic equilibrium analysis, Katz–Bonacich centrality, and mechanism design theory, they demonstrate that optimal evaluation rules generally deviate from true output—employing negative matching with low weights when effort dominates, and positive matching with high weights when advantages dominate. The work also provides the first quantification of efficiency loss due to pairwise stability constraints and characterizes the intrinsic relationship between equilibrium effort, network position, and effective advantage, offering tailored optimal strategies for diverse production environments.

This study addresses inequities and inefficiencies in computational resource allocation within federated digital research infrastructures (DRIs) by framing resource scheduling as an intersection of technical and policy considerations. Integrating policy analysis, mechanism design, and simulation experiments, the work reveals a significant disconnect between resource allocation and actual utilization. It innovatively applies Braess’s paradox to demonstrate how uncoordinated federation can exacerbate load imbalances. Leveraging algorithmic game theory, transport economics models, and simulations based on both real-world (Fresco/Anvil) and synthetic workloads—complemented by empirical investigation of international DRI allocation mechanisms—the research finds that carefully tuned, transparent scheduling heuristics can closely approximate ideal performance. While federated approaches hold promise, they require coordinated governance, and current systems notably lack effective evaluation of scientific value generated per unit of allocated resources.

This work investigates the performance trade-offs of non-adaptive strategies in stochastic load balancing. It proposes a two-stage model: in the first stage, each job reserves up to $k$ machines based on the task size distribution; in the second stage, after observing the actual job size, it is assigned to one of the reserved machines to minimize the expected makespan. The paper establishes, for the first time in this setting, a “power of two choices” theory, showing that under identical machines, reserving just two machines per job suffices to achieve a constant-factor approximation to the omniscient optimal solution. For related machines, it provides an $O(\log m / \log \log m)$-approximation and a bicriteria constant-factor approximation, and further proves that with 2-reservation, one can approximate the adaptively optimal solution.

This work addresses the challenge of reconciling Pareto efficiency with max-min fairness in machine learning training resource allocation under heterogeneous user workload valuations. The authors propose and deploy a market-based GPU quota exchange platform featuring a dynamic pricing mechanism that enables users to explicitly express the value of their tasks, thereby guiding efficient resource allocation. To the best of our knowledge, this system is the first to simultaneously guarantee both Pareto efficiency and max-min fairness in settings with heterogeneous task values, successfully integrating market mechanisms into a large-scale production environment. Deployment within Google demonstrates significant improvements in resource utilization, effective support for diverse business-critical workloads, and the opening of new avenues for scheduling optimization.

This study addresses the allocation of multiple heterogeneous resource types in hierarchical organizations, where certain resources can be transformed into one another. The problem is formulated as a market equilibrium model incorporating structural constraints inherent to the hierarchy. To solve it efficiently, the authors propose a novel two-stage approximation algorithm: first solving a tractable instance that respects the hierarchical structure, then iteratively refining the solution to handle general cases. This work introduces, for the first time, a two-stage approximation framework to hierarchical resource allocation with conversion capabilities, establishing both the guaranteed existence of feasible equilibria and computational efficiency. Experiments on real-world Google TPU/GPU allocation datasets demonstrate the algorithm’s effectiveness and rapid convergence.