arXiv:2610. 21001v1 Announce Type: new Abstract: Column generation folklore warns that initial columns hurt: a warm start may bias the dual and slow the root of branch-price-and-cut.
arXiv:2607. 13373v1 Announce Type: cross Abstract: Column generation (CG) is central to many large-scale optimization algorithms, including branch-price-and-cut methods for vehicle routing problems, but unstable dual solutions can substantially slow its convergence.
The paper introduces CostStart, a method that learns warm starts for the simplex algorithm by predicting which variables will be basic at optimality and then selecting the number of predicted columns, algorithm, and pricing rule based on measured instruction counts. Experiments on 25 classes of linear programs show that while a strong supervised predictor can reduce dual simplex iterations by two thirds, it only cuts solve time by one third; CostStart mitigates this gap and roughly halves solve time compared to the solver’s default start, outperforming two published methods on unseen problems up to four times larger.
arXiv:2610.14792v1 Announce Type: new Abstract: Online, same-day delivery requires operators to constantly update their routes, and the routing cost for deliveries increases with the number of custom...
arXiv:2607. 20594v1 Announce Type: cross Abstract: When does a weight-tied looped transformer -- one block applied T times -- implement an actual algorithm?
The paper presents a method for reducing the size of Quadratic Unconstrained Binary Optimization (QUBO) models used to solve the Capacitated Vehicle Routing Problem with Time Windows (CVRPTW) on quantum annealers. It introduces adaptive penalty calibration to improve constraint satisfaction and replaces hand‑tuned merge heuristics with a graph neural network (GNN) that consistently achieves higher feasibility across Solomon benchmark families. Experiments on simulated annealing and a D‑Wave Advantage2 processor show significant reductions in constraint violations and improved feasibility rates, with the QUBO size remaining 5–6 times smaller.
arXiv:2608. 04804v1 Announce Type: cross Abstract: Frontier language models can resolve repository-level software issues, but each attempt is expensive, and existing routers select a model from the issue text alone.
The paper evaluates a learned request‑routing policy for disaggregated large‑language‑model serving, where compute‑heavy prefill and memory‑heavy decode stages run on separate GPU pools. Using a discrete‑event simulator and real NVIDIA A40 GPUs, the calibrated router—leveraging prompt length, predicted output length, KV‑cache pressure, and SLO class—outperforms round‑robin, least‑loaded, and length‑based heuristics, achieving the highest mean goodput (0.864) and lowest variance across three mixed, bursty arrival traces. Hardware calibration proves critical, providing a 4.5‑point goodput boost and roughly 40 % of the tail‑latency advantage, and the learned router can match round‑robin performance with one fewer GPU in certain scenarios.
arXiv:2608. 08853v1 Announce Type: new Abstract: Sparse Mixture-of-Experts (MoE) routers commonly use the same scores both to select experts and to weight their already-computed outputs.
arXiv:2608. 19953v1 Announce Type: new Abstract: Mixed-Integer Linear Programming (MILP) is a fundamental problem class in operations research and combinatorial optimization, with broad applications to industrial decision-making.
The paper introduces Temperon, a training strategy that uses plain SGD for the first 43% of the epoch budget and then hands off to a SAM‑wrapped Muon refiner for the remaining training. On datasets such as CIFAR‑10/100, SVHN, and Tiny ImageNet, Temperon achieves the same or better accuracy as full‑time SAM while reaching key performance targets faster and at lower cost. Ablation studies show that the Muon refiner contributes the majority of the performance gain, while the initial SGD explorer and its restarts add negligible benefit.
arXiv:2609.22951v1 Announce Type: cross Abstract: Enterprise agentic systems that route every trajectory step to a frontier model waste 60-80% of their inference budget on subtasks that smaller model...
Large traveling salesman problem (TSP) instances require a solver to allocate limited computation while preserving the validity of its outputs. Existing neural--operations-research (OR) hybrids predict guidance without requiring learned transitions to satisfy constraints discovered during search.