Density-Peak Initialized K-Means and Multi-Objective Integer Optimization for Dynamic Collaborative Resource Allocation

Dynamic collaborative organizations must allocate limited decision-making resources under task uncertainty, cross-role workload coupling and reliability constraints. This study develops an integrated type-determination and quantity-optimization framework: task features are weighted by entropy and partitioned by a deterministic density-peak initialized K-means step, and a cost-efficiency-reliability integer program, solved by an adaptive-crossover genetic algorithm and verified against exhaustive enumeration, determines the number of seats of each type. The four seat types are fixed by the domain requirement for four command roles rather than derived from the cluster labels. Under the documented assumptions of a 6 h decision cycle and c = g = (4, 5, 6, 3) simulation units, the exact optimum is [2, 3, 2, 3], with cost 44, efficiency 0.8040 and system reliability 0.9700; it lies on the exact Pareto front of the 64 feasible configurations, and one hundred independent GA runs recover it. On synthetic data with known structure the clustering step recovers the true partition as accurately as random or k-means++ initialization (mean adjusted Rand index 0.973) and returns the same partition on every run. A simplified rolling-horizon simulation with sequential arrivals shows that rolling reconfiguration reduces backlog at moderate demand. Sensitivity analyses show that the optimum is unchanged under one-step changes in the AHP judgments and under most joint perturbations of the cost coefficients, but it depends on the stated priority order and, near weight ties, on the exact intensities of the judgments; resource and failure-rate shocks can make the problem infeasible, which is reported rather than replaced with unsupported values.

Authors

Institutions

Publication Details

Journal
Applied Sciences
Published
2026-11-13
DOI
https://doi.org/10.3390/app162211277
Primary Topic
Military Defense Systems Analysis
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Density-Peak Initialized K-Means and Multi-Objective Integer Optimization for Dynamic Collaborative Resource Allocation

Longyue Li, Guoqing Zhang, Huizhen Zhao, Panrong wang
Applied Sciences
Military Defense Systems Analysis
article

Density-Peak Initialized K-Means and Multi-Objective Integer Optimization for Dynamic Collaborative Resource Allocation

Longyue Li, Guoqing Zhang, Huizhen Zhao, Panrong wang
article en

Abstract

Dynamic collaborative organizations must allocate limited decision-making resources under task uncertainty, cross-role workload coupling and reliability constraints. This study develops an integrated type-determination and quantity-optimization framework: task features are weighted by entropy and partitioned by a deterministic density-peak initialized K-means step, and a cost-efficiency-reliability integer program, solved by an adaptive-crossover genetic algorithm and verified against exhaustive enumeration, determines the number of seats of each type. The four seat types are fixed by the domain requirement for four command roles rather than derived from the cluster labels. Under the documented assumptions of a 6 h decision cycle and c = g = (4, 5, 6, 3) simulation units, the exact optimum is [2, 3, 2, 3], with cost 44, efficiency 0.8040 and system reliability 0.9700; it lies on the exact Pareto front of the 64 feasible configurations, and one hundred independent GA runs recover it. On synthetic data with known structure the clustering step recovers the true partition as accurately as random or k-means++ initialization (mean adjusted Rand index 0.973) and returns the same partition on every run. A simplified rolling-horizon simulation with sequential arrivals shows that rolling reconfiguration reduces backlog at moderate demand. Sensitivity analyses show that the optimum is unchanged under one-step changes in the AHP judgments and under most joint perturbations of the cost coefficients, but it depends on the stated priority order and, near weight ties, on the exact intensities of the judgments; resource and failure-rate shocks can make the problem infeasible, which is reported rather than replaced with unsupported values.

Applied SciencesVol. 16(22)
Air Force Engineering University (CN)
Natural Science Basic Research Program of Shaanxi Province
Industry, innovation and infrastructure
Military Defense Systems Analysis
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