The Optimal Competitive Ratio Is 19/16 for Semi-Online Scheduling on Five Identical Machines with Nonincreasing Processing Times

Jobs arrive one at a time in nonincreasing order of processing time and must be assigned immediately to five identical machines. The input may stop after any job. We give a deterministic algorithm whose makespan on every prefix is at most 19/16 times its offline optimum. The algorithm tries placements using the sizes of the jobs already seen and always uses a least-loaded machine from job 13 onward. Before a first violation, every machine holds at most three jobs, so it suffices to examine the first sixteen jobs. Depending on whether an optimal packing requires a four-job bin, the available slots force particular machines to receive jobs 6 and 7 or force job 8 to fail its first trial. Comparing these necessary placements with the algorithm excludes the remaining stopping points. An input with three job sizes and its early stops give a matching lower bound.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-21
DOI
https://doi.org/10.5281/zenodo.24409951
Primary Topic
Optimization and Search Problems
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

The Optimal Competitive Ratio Is 19/16 for Semi-Online Scheduling on Five Identical Machines with Nonincreasing Processing Times

Chaojin Wu
Zenodo (CERN European Organization for Nuclear Research)
Optimization and Search Problems
preprint

The Optimal Competitive Ratio Is 19/16 for Semi-Online Scheduling on Five Identical Machines with Nonincreasing Processing Times

Chaojin Wu
preprint en

Abstract

Jobs arrive one at a time in nonincreasing order of processing time and must be assigned immediately to five identical machines. The input may stop after any job. We give a deterministic algorithm whose makespan on every prefix is at most 19/16 times its offline optimum. The algorithm tries placements using the sizes of the jobs already seen and always uses a least-loaded machine from job 13 onward. Before a first violation, every machine holds at most three jobs, so it suffices to examine the first sixteen jobs. Depending on whether an optimal packing requires a four-job bin, the available slots force particular machines to receive jobs 6 and 7 or force job 8 to fail its first trial. Comparing these necessary placements with the algorithm excludes the remaining stopping points. An input with three job sizes and its early stops give a matching lower bound.

Zenodo (CERN European Organization for Nuclear Research)
Optimization and Search Problems
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon