Synchronization-externality-aware scheduling for autonomous RoboVan-drone collaborative delivery
Year: 2026
Type: preprint
Abstract: Autonomous RoboVan-drone delivery couples aerial assignments through shared launch, recovery, charging, and moving-rendezvous resources. We propose Predictive Synchronization-Externality-Aware Rolling Online Scheduling (P-SEAROS), which separates a focal order's direct cost from the downstream synchronization burden imposed on the remaining schedule. Candidate scores are refreshed after each accepted insertion. On held-out data, the externality estimator reaches a Spearman correlation of 0.612 and a raw MAE of 4.74 cost units. Across 20 paired scheduling seeds, P-SEAROS reduces mean system cost by 17.2%, 17.0%, and 14.3% at 80, 100, and 120 requests, with Holm-adjusted two-sided p-values below 0.01. A 20-seed mechanism comparison shows that externality-aware scoring accounts for most of the gain, while prediction provides a smaller increment and exact predicted marginal scoring improves cost by only about 1.2–1.3%. The 95th-percentile runtime remains below 30 ms per order for both approximate and exact predicted scoring at the tested scale. Descriptive regime analyses suggest larger benefits when shared-resource coupling materially affects downstream schedules.
Source: Research Square
Author: Chen Li
Institution: Wuhan Institute of Technology
Language: English
Subfield: Industrial and Manufacturing Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7214738126