UAV-UGV Cooperative Trajectory Optimization and Task Allocation for Medical Rescue Tasks in Post-Disaster Environments

Kaiyuan Chen; Shuo Wang; Wannian Liang; Wanpeng Zhao; Yongxi Liu; Yuanqing Xia

REVIEW 4 major objections 5 minor 27 references

UAV-UGV Cooperative Trajectory Optimization and Task Allocation for Medical Rescue Tasks in Post-Disaster Environments

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read A UAV-UGV rescue planner cuts 15-task medical missions to 26.7 minutes.

desk verdict A routine integration of standard methods with an experimental section that is clearly unfinished: future-tense results, contradictory numbers, and no statistical evidence. read the letter →

arxiv 2506.06136 v2 pith:TOV3DA4G submitted 2025-06-06 cs.RO cs.MA

Kaiyuan Chen , Wanpeng Zhao , Yongxi Liu , Yuanqing Xia , Wannian Liang , Shuo Wang This is my paper

classification cs.ROcs.MA

keywords UAV-UGVcooperationpost-disasterrescuetaskallocationgeneticalgorithmInformed-RRT*CMA-EStrajectoryoptimizationmedicallogistics

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper sets out to show that a heterogeneous team of aerial and ground robots can run post-disaster medical rescue missions far faster than conventional allocation schemes if task assignment, path generation, and route sequencing are optimized together. The proposed framework chains an enhanced genetic algorithm for task allocation, Informed-RRT* for collision-free path generation, and CMA-ES for refining the order in which each vehicle visits its tasks. In a simulated 20 km by 20 km post-earthquake environment with 15 task points, the authors report a total mission completion time of 26.7 minutes, beating K-Means clustering and random allocation by more than 73 percent, with CMA-ES cutting total traveled distance by 15.1 percent (20.3 percent in the figure). The authors also report that the approach scales to 60 tasks and 30 vehicles with no loss of efficiency. The significance, if the numbers hold, is a practical template for coordinating drone-and-ground-vehicle teams in time-critical disaster response.

What carries the argument

The load-bearing mechanism is the three-stage pipeline itself. The Enhanced Genetic Algorithm encodes each candidate solution as a chromosome in which tasks are grouped by vehicle, uses tournament selection, partially matched crossover with repair, an adaptive mutation probability that decays over generations, and elite preservation to allocate tasks while respecting range and payload limits. Informed-RRT* then restricts sampling to an ellipsoid defined by the current best path cost, generating collision-free waypoint paths that connect the base, the assigned tasks, and the return leg. CMA-ES, operating on real-valued vectors whose element rankings define the visiting permutation, adapts its covariance matrix to learn correlations among task points, so that the refined sequence reduces redundant travel. The decomposition is what makes the combined problem tractable: global allocation feeds local planning, and local planning feeds sequence refinement.

What would settle it

Run the simulation under the stated setup—20 km × 20 km map, five circular obstacles of 500 m to 1.5 km diameter, ten UAVs (60 km/h, 15 km range, 5 kg payload) and five UGVs (30 km/h, 25 km range, 50 kg payload), base at (0,0), and the same 15 task coordinates—and check whether the reported 26.7-minute completion time, the 73% margin over K-Means/random, and the 15.1% or 20.3% CMA-ES distance reduction are reproduced.

Watch

Extended reading notes

Core claim

The paper's central claim is that the 'Allocate-Plan-Optimize' decomposition—EGA for global task allocation, Informed-RRT* for locally feasible collision-free trajectories, and CMA-ES for per-vehicle task sequencing—produces mission plans that are both feasible and substantially more efficient than the baselines it is compared against. Specifically, the authors claim that for a 15-task scenario with ten UAVs and five UGVs, the framework completes the mission in 26.7 minutes whereas K-Means clustering and random allocation each take 100.0 minutes; that doubling the task count to 30 raises the optimized time only to 39.8 minutes; and that increasing to 60 tasks with 20 UAVs and 10 UGVs yields 35.0 minutes, evidence of parallelization gains. The paper further claims that optimizing the visiting sequence with CMA-ES reduces total path length, reported as a 15.1 percent improvement in the abstract and a 20.3 percent improvement (61.2 km to 48.8 km) in Figure 5, and that all generated trajectories respect obstacle boundaries and vehicle endurance constraints.

Load-bearing premise

The whole quantitative case rests on whether the headline numbers are actual results of the described pipeline rather than planned or estimated figures, since the results section is written in future tense and the distance reduction is given as 15.1% in the abstract but 20.3% in Figure 5.

Editorial extensions

If this is right

If the reported numbers hold, rescue planners can expect a GA-based allocator followed by Informed-RRT* and CMA-ES to turn a 15-task medical delivery into a 26.7-minute operation, a 73% reduction over clustering or random assignment.
The scalability results imply that adding vehicles and tasks in proportion keeps mission time roughly constant, so the framework would remain useful as disaster zones grow.
The CMA-ES sequencing step alone is credited with a 15-20% reduction in total traveled distance, meaning route reordering is a significant lever independent of the allocator.
Because the framework decomposes allocation, pathfinding, and sequencing, each component can be tuned or replaced without redesigning the whole system, which is a direct corollary of the proposed architecture.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

A natural testable extension is to swap the Informed-RRT* module for a kinodynamic planner and check whether the same 73% margin survives under acceleration and turning constraints, which the paper does not simulate.
The continuous-relaxation trick used to let CMA-ES optimize discrete task permutations is transferable to other routing and scheduling problems in disaster logistics, not just UAV-UGV teams.
The static-environment assumption means the headline numbers are an upper bound on real-world performance; introducing dynamic replanning would likely erode some of the margin, but the relative ordering of methods may persist.
Since the paper treats energy as proportional to path length, the reported distance savings imply proportional battery savings; measuring with a nonlinear energy model would show whether that inference holds.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

The paper is not ready for publication. The results section is written in the future tense and the few numbers it reports are mutually inconsistent, so the headline claims of a 26.7-minute completion time and a 73% improvement are unsupported.

What is actually new: not much scientifically. The pipeline—EGA for allocation, Informed-RRT* for paths, CMA-ES for sequence refinement—is a straightforward combination of off-the-shelf methods. That said, the integrated application to heterogeneous UAV-UGV medical delivery is a sensible engineering idea, and the problem formulation in Section 3 is clear. The authors also deserve credit for an honest limitations paragraph (static environment, no dynamic deconfliction, abstracted energy model).

The soft spots are load-bearing. Section 5.2 uses conditional language: 'Expected results ... will demonstrate', 'Fig. 6 ... would visually depict', 'The analysis will confirm'. This reads like a template, not a completed study. The concrete numbers contradict each other: the abstract claims a 15.1% distance reduction; the caption of Fig. 5 claims 20.3% (61.2 to 48.8 km). Figure 3 shows EGA producing 139.9 km of total path, while Fig. 5 reports 61.2 km before CMA-ES in what should be the same 15-task scenario. Table 1 lists identical 100.0-minute baselines for K-Means and random allocation, which is physically implausible unless the baselines are degenerate. There are no seeds, repetitions, or statistical tests, and the baseline implementations are not specified.

There is also a citation-inflation problem: references 13–26 are a block of commodity-price and house-price forecasting papers that have nothing to do with multi-robot planning. That is not a technical flaw, but it undermines confidence in the authors' scholarly care.

The paper does promise code on GitHub, which is a positive sign, but the promise alone does not validate the numbers.

Bottom line: a serious referee would reject this in its current form. The underlying framework might work, but the evidence is not there. The authors should run the experiments, report consistent numbers, add proper baselines and error bars, and cut the irrelevant references. If they do that, the paper could become a modest engineering contribution to the niche disaster-rescue community. As it stands, it deserves a desk reject, and I would not cite it.

Referee Report

4 major / 5 minor

Summary. The paper proposes an integrated framework for cooperative task allocation and trajectory optimization in heterogeneous UAV-UGV teams for medical rescue in post-disaster environments. The framework combines an Enhanced Genetic Algorithm (EGA) for task allocation, Informed-RRT* for collision-free path generation, and CMA-ES for task-sequencing and trajectory refinement. The authors report simulation results claiming a 26.7-minute mission completion time for a 15-task scenario, over 73% improvement versus K-Means and random allocation, and a 15.1% (Abstract) or 20.3% (Fig. 5) reduction in total traveled distance after CMA-ES optimization. However, Section 5.2 is written in future/conditional tense, the quantitative results are internally inconsistent, and no complete experimental protocol (task coordinates, seeds, repetitions, parameter values, baseline details) is provided.

Significance. If validated, the proposed modular pipeline (EGA, Informed-RRT*, CMA-ES) addresses a practically important problem, and the integration itself is a reasonable design. The paper also provides a GitHub link for code availability, which is a positive step. However, the central quantitative claims—the mission-completion-time reductions and distance savings—are not supported by a completed, consistent experimental record. The results section describes anticipated rather than actual outcomes, and the few explicit numbers contradict one another. As such, the significance of the contribution cannot currently be assessed. The contribution may have merit as an algorithmic proposal, but the paper does not present the evidence needed to establish that merit.

major comments (4)

[Section 5.2] Section 5.2 is written in future/conditional tense and does not report completed experiments. The text states: 'Expected results, visualized using bar charts Fig. 3, will demonstrate...', 'Fig. 6 ... would visually depict', and 'The analysis will confirm...'. These phrases indicate that the simulation results are anticipated rather than actually obtained. The quantitative claims in the Abstract (26.7 min, >73% improvement, 15.1% distance reduction) and Conclusion therefore rest on results that are not presented as actual outcomes. This is a load-bearing flaw: the central contribution of the paper is the claimed performance advantage, and without valid experimental outputs the claim is unsupported.
[Abstract vs. Fig. 5 and Fig. 3] The reported quantitative results are internally inconsistent. The Abstract states a '15.1% reduction in total traveled distance after CMA-ES optimization,' while the Fig. 5 caption reports a reduction from 61.2 km to 48.8 km, i.e., 20.3%. Fig. 3 reports the EGA total path length as 139.9 km for the same 15-task scenario, while Fig. 5 gives 61.2 km 'before optimization.' These numbers cannot all describe the same scenario. Table 1 lists identical baseline times of 100.0 min for both K-Means and Random Allocation at 15 tasks, which is implausible for two different methods. The reader cannot determine which, if any, of these numbers are actual outputs of the pipeline.
[Section 5.1 and Table 1] The experimental protocol is not described sufficiently for reproducibility or for assessing the claimed gains. No task coordinates, obstacle positions, random seeds, number of repetitions, or statistical variability are given. The algorithm parameters for EGA and CMA-ES are described only qualitatively ('set according to the specific scale and complexity of the problem'), with no numerical values. The baseline implementations of K-Means and random allocation are not specified beyond their names. Without this information, the reported improvements cannot be verified or meaningfully compared with other methods.
[Section 3.4 and Table 1] The reported mission completion times are not derived from the stated vehicle speeds and path lengths. Eq. (3) defines makespan as the maximum per-vehicle mission time, but the paper does not provide per-vehicle tour lengths, the speed profile used to convert path length to time, or a calculation that yields 26.7 min for the 15-task scenario. The total path length of 48.8 km after optimization and the vehicle speeds (UAV 60 km/h, UGV 30 km/h) do not by themselves imply any particular makespan. The claim is therefore unverifiable.

minor comments (5)

[Equations (9) and (11)] Equations (9) and (11) are identical; the duplicate presentation of the Informed-RRT* sampling domain should be removed.
[Sections 2.2 and 2.3] The literature review includes many references on commodity price forecasting and machine learning (refs. [13]–[26]) that are not related to task allocation or trajectory planning; these should either be removed or explicitly connected to the topic.
[Section 5.2] Section 5.2 contains incomplete sentences, e.g., 'In contrast, K-Means and Random, both 100 min in the baseline.' This should be rewritten for clarity and grammatical completeness.
[Entire manuscript] The manuscript has frequent formatting errors with spacing, e.g., 'UA V', 'T ask', 'T rajectory' throughout, and should be carefully proofread.
[Figure 6] Fig. 6 is described conditionally ('would visually depict'), which is not appropriate for a results section; if the figure is included, it should be described in past tense with actual observations.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the reported gains are direct optimization outputs on the stated objectives, not fitted inputs renamed as predictions, and no load-bearing self-citation or imported uniqueness claim is present.

full rationale

The paper is an engineering optimization pipeline (EGA allocation, Informed-RRT* path generation, CMA-ES sequence refinement) evaluated by simulation, not a derivation whose conclusion is equivalent to its premises. The objectives minimized — Eq. (5) for EGA task allocation, Eq. (12) for CMA-ES visiting-order optimization, and Eq. (3) for makespan — are the same path-length and mission-time metrics subsequently reported as improvements; that is direct optimization of the stated objective, not a prediction that reduces by construction to its inputs. No fitted parameter is renamed as a prediction: the algorithm parameters in Section 5.1 are configured heuristically and are not fit to the reported 26.7-minute, 73%, or 15.1% outcomes. The reference list is entirely external; there is no self-citation chain, no imported uniqueness theorem, no ansatz smuggled in via prior work by the same authors, and no renaming of a known empirical pattern as a new organizing principle. The manuscript does have serious evidential and consistency problems — Section 5.2 uses future/conditional language ("Expected results, visualized using bar charts Fig. 3, will demonstrate..."), Fig. 3 and Fig. 5 give mutually inconsistent total path lengths for the same 15-task pipeline, Table 1 lists identical 100.0-minute baselines for K-Means and Random at 15 tasks, and no seeds or repetition protocols are reported. These are completeness, reproducibility, and internal-consistency flaws, not circularity. The central claim is unsupported on the evidence presented, but it is not circular; no circular step can be exhibited with the required specificity, so the circularity score is 0.

Assumptions & free parameters 2 free parameters · 5 assumptions · 0 invented entities

The paper introduces no new physical or mathematical entities. Its assumptions are standard simplifications for a planning simulation, but they are not validated against real-world data. The main free parameters are the hyperparameters of the EGA and CMA-ES, which are not reported with specific values.

free parameters (2)

EGA population size, generations, elite ratio, initial mutation rate
Algorithm 1 lists these as inputs but no values are reported in Section 5.1; they are hand-chosen and affect results.
CMA-ES initial mean, covariance, population size
Described as 'carefully selected' without specific values in Section 5.1.

assumptions (5)

domain assumption The environment is modeled as a static 2D Euclidean space with known obstacles and fixed task locations (Section 3.1).
Assumes no dynamic changes during the mission, acknowledged by authors in Section 5.2.1.
domain assumption All vehicles have perfect communication and idealized control (Section 3.2 and Section 5.2.1).
The paper assumes no communication delays or packet loss, which are prevalent in disaster zones.
standard math Informed-RRT* provides asymptotic optimality under certain conditions (Section 4.2).
The paper relies on this property to claim near-optimal trajectories, but does not specify the conditions.
domain assumption Tasks are independent with no precedence constraints or time windows (Section 3.3).
The paper simplifies the problem by ignoring task dependencies and priorities.
domain assumption Euclidean distance is used as the cost metric and proxy for energy/time (Equations 1, 2, 5, 12).
The paper acknowledges that actual energy depends on trajectory curvature and dynamics, but uses straight-line distances.

how reviews work

0 comments

Cite this review

Pith. "Pith review of UAV-UGV Cooperative Trajectory Optimization and Task Allocation for Medical Rescue Tasks in Post-Disaster Environments." pith.science (2026). https://pith.science/paper/TOV3DA4G

@misc{pith2026250606136,
  author       = {Pith},
  title        = {Pith review of: UAV-UGV Cooperative Trajectory Optimization and Task Allocation for Medical Rescue Tasks in Post-Disaster Environments},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/TOV3DA4G}},
  note         = {Machine review of arXiv:2506.06136}
}

read the original abstract

In post-disaster scenarios, rapid and efficient delivery of medical resources is critical and challenging due to severe damage to infrastructure. To provide an optimized solution, we propose a cooperative trajectory optimization and task allocation framework leveraging unmanned aerial vehicles (UAVs) and unmanned ground vehicles (UGVs). This study integrates a Genetic Algorithm (GA) for efficient task allocation among multiple UAVs and UGVs, and employs an informed-RRT* (Rapidly-exploring Random Tree Star) algorithm for collision-free trajectory generation. Further optimization of task sequencing and path efficiency is conducted using Covariance Matrix Adaptation Evolution Strategy (CMA-ES). Simulation experiments conducted in a realistic post-disaster environment demonstrate that our proposed approach significantly improves the overall efficiency of medical rescue operations compared to traditional strategies. Specifically, our method reduces the total mission completion time to 26.7 minutes for a 15-task scenario, outperforming K-Means clustering and random allocation by over 73%. Furthermore, the framework achieves a substantial 15.1% reduction in total traveled distance after CMA-ES optimization. The cooperative utilization of UAVs and UGVs effectively balances their complementary advantages, highlighting the system's scalability and practicality for real-world deployment.

Figures

Figures reproduced from arXiv: 2506.06136 by the authors.

**Figure 2.** Integrated framework diagram The strength of this framework lies in its structured integration of specialized algorithms. The EGA provides a robust global task distribution considering vehicle heterogeneity. Its output (task assignments X) directly defines the subproblems solved by the subsequent path planning and sequence optimization stages for each vehicle. Informed-RRT* acts as a crucial feasibility engine, e… view at source ↗

**Figure 3.** Comparison of total path length generated by dif [PITH_FULL_IMAGE:figures/full_fig_p010_3.png] view at source ↗

Figures from the paper (3 more)

**Figure 4.** Figure 4: Convergence curve of the EGA for task allocation. [PITH_FULL_IMAGE:figures/full_fig_p011_4.png]

**Figure 7.** Figure 7: Final cooperative mission plan for the heterogeneous [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]

**Figure 6.** Figure 6: Visual comparison of a representative vehicle tra [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

27 extracted references · 26 canonical work pages

[1]

S ¸enol Balaban, C

M. S ¸enol Balaban, C. Do˘ gulu, N. Akdede, H. Ako˘ glu, O. Karakayalı, S. Yılmaz, S. Yılmaz, T. Ajobiewe, S. G¨ uzel, G.˙Ikizer, et al. , “Emergency response, and community impact after february 6, 2023 kahraman- mara¸ s pazarcık and elbistan earthquakes: reconnais- sance findings and observations on affected region in t¨ urkiye,”Bulletin of Earthquake E...

work page 2023
[2]

Heterogeneous multi-robot collaboration for coverage path planning in partially known dynamic environments,

G. G. R. de Castro, T. M. B. Santos, F. A. A. An- drade, J. Lima, D. B. Haddad, L. d. M. Hon´ orio, and M. F. Pinto, “Heterogeneous multi-robot collaboration for coverage path planning in partially known dynamic environments,” Machines, vol. 12, no. 3, 2024

work page 2024
[3]

Cooperative aerial-ground multi- robot system for automated construction tasks,

M. Krizmancic, B. Arbanas, T. Petrovic, F. Petric, and S. Bogdan, “Cooperative aerial-ground multi- robot system for automated construction tasks,” IEEE Robotics and Automation Letters , vol. 5, no. 2, pp. 798–805, 2020

work page 2020
[4]

Cooperative control of a heterogeneous multi- robot system based on relative localization,

M. Cognetti, G. Oriolo, P. Peliti, L. Rosa, and P. Ste- gagno, “Cooperative control of a heterogeneous multi- robot system based on relative localization,” in 2014 September 12, 2025 7:44 ws-us 14 REFERENCES IEEE/RSJ International Conference on Intelligent Robots and Systems , pp. 350–356, 2014

work page 2014
[5]

Collaborative multi-robot search and rescue: Planning, coordination, perception, and active vision,

J. P. Queralta, J. Taipalmaa, B. Can Pullinen, V. K. Sarker, T. Nguyen Gia, H. Tenhunen, M. Gabbouj, J. Raitoharju, and T. Westerlund, “Collaborative multi-robot search and rescue: Planning, coordination, perception, and active vision,” IEEE Access, vol. 8, pp. 191617–191643, 2020

work page 2020
[6]

Task assignment of multi- robot systems based on improved genetic algorithms,

S. Li, X. Xu, and L. Zuo, “Task assignment of multi- robot systems based on improved genetic algorithms,” in 2015 IEEE International Conference on Mechatron- ics and Automation (ICMA) , pp. 1430–1435, 2015

work page 2015
[7]

A pso-based multi-robot cooperation method for target searching in unknown environments,

M. Dadgar, S. Jafari, and A. Hamzeh, “A pso-based multi-robot cooperation method for target searching in unknown environments,” Neurocomputing, vol. 177, pp. 62–74, 2016

2016
[8]

Bal- ancing task allocation in multi-robot systems using k-means clustering and auction based mechanisms,

M. Elango, S. Nachiappan, and M. K. Tiwari, “Bal- ancing task allocation in multi-robot systems using k-means clustering and auction based mechanisms,” Expert Systems with Applications , vol. 38, no. 6, pp. 6486–6491, 2011

work page 2011

Show all 27 references

[9]

Multi robot distance based formation using parallel genetic algorithm,

A. L´ opez-Gonz´ alez, J. Meda Campa˜ na, E. Hern´ andez Mart ´ ınez, and P. P. Contro, “Multi robot distance based formation using parallel genetic algorithm,” Ap- plied Soft Computing , vol. 86, p. 105929, 2020

2020
[10]

Multi-robot path planning based on the developed rrt* algorithm,

Y. Li, R. Cui, C. Yang, and D. Xu, “Multi-robot path planning based on the developed rrt* algorithm,” in Proceedings of the 32nd Chinese Control Conference , pp. 7049–7053, 2013

2013
[11]

A novel hybrid framework for single and multi-robot path planning in a complex industrial environment,

S. Kumar and A. Sikander, “A novel hybrid framework for single and multi-robot path planning in a complex industrial environment,” Journal of Intelligent Manu- facturing, vol. 35, no. 2, pp. 587–612, 2024

2024
[12]

Deep reinforcement learning for swarm navigation using different evolution strate- gies,

´A. D ´ ıez and F. Aznar, “Deep reinforcement learning for swarm navigation using different evolution strate- gies,” in Proceedings of the 2024 8th International Conference on Intelligent Systems, Metaheuristics & Swarm Intelligence , pp. 14–19, 2024

2024
[13]

Commodity price fore- casting via neural networks for coffee, corn, cotton, oats, soybeans, soybean oil, sugar, and wheat

Xu, Xiaojie and Zhang, Yun. Commodity price fore- casting via neural networks for coffee, corn, cotton, oats, soybeans, soybean oil, sugar, and wheat. Intel- ligent Systems in Accounting, Finance and Manage- ment, 29(3):169–181, 2022

2022
[14]

Second-hand house price index forecasting with neural networks

Xu, Xiaojie and Zhang, Yun. Second-hand house price index forecasting with neural networks. Journal of Property Research, 39:1–22, 2021

2021
[15]

Soybean and Soybean Oil Price Forecasting through the Nonlinear Autore- gressive Neural Network (NARNN) and NARNN with Exogenous Inputs (NARNN–X)

Xu, Xiaojie and Zhang, Yun. Soybean and Soybean Oil Price Forecasting through the Nonlinear Autore- gressive Neural Network (NARNN) and NARNN with Exogenous Inputs (NARNN–X). Intelligent Systems with Applications, 13:200061, 2022

2022
[16]

Thermal coal futures trading volume predictions through the neural network

Jin, Bingzi and Xu, Xiaojie and Zhang, Yun. Thermal coal futures trading volume predictions through the neural network. Journal of Modelling in Management , 20: , 2024

2024
[17]

Peanut oil price change forecasts through the neural network

Jin, Bingzi and Xu, Xiaojie and Zhang, Yun. Peanut oil price change forecasts through the neural network. Foresight, 27: , 2025

2025
[18]

Corn cash price forecast- ing with neural networks

Xu, Xiaojie and Zhang, Yun. Corn cash price forecast- ing with neural networks. Computers and Electronics in Agriculture, 184:106120, 2021

2021
[19]

House price forecasting with neural networks

Xu, Xiaojie and Zhang, Yun. House price forecasting with neural networks. Intelligent Systems with Appli- cations, 12:200052, 2021

2021
[20]

Contemporaneous and Granger causality among US corn cash and futures prices

Xu, Xiaojie. Contemporaneous and Granger causality among US corn cash and futures prices. European Re- view of Agricultural Economics , 46(4):663–695, 2019

2019
[21]

Predictions of residential property price indices for China via machine learning models

Jin, Bingzi and Xu, Xiaojie. Predictions of residential property price indices for China via machine learning models. Quality & Quantity , 59:1481–1513, 2025

2025
[22]

Predictions of steel price indices through machine learning for the regional northeast Chinese market

Jin, Bingzi and Xu, Xiaojie. Predictions of steel price indices through machine learning for the regional northeast Chinese market. Neural Computing and Ap- plications, 36:20863–20882, 2024

2024
[23]

Price forecasts of ten steel products using Gaussian process regres- sions

Xu, Xiaojie and Zhang, Yun. Price forecasts of ten steel products using Gaussian process regres- sions. Engineering Applications of Artificial Intelli- gence, 126:106870, 2023

2023
[24]

Contemporaneous causal orderings among prices of retail properties: evidence from Chinese cities through vector error-correction modeling and directed acyclic graphs

Jin, Bingzi and Xu, Xiaojie. Contemporaneous causal orderings among prices of retail properties: evidence from Chinese cities through vector error-correction modeling and directed acyclic graphs. Journal of Fi- nancial Management of Property and Construction , 2025

2025
[25]

Contemporaneous causal- ity among price indices of ten major steel products

Jin, Bingzi and Xu, Xiaojie. Contemporaneous causal- ity among price indices of ten major steel products. Ironmaking & Steelmaking , 51: , 2024

2024
[26]

Palladium Price Predic- tions via Machine Learning

Jin, Bingzi and Xu, Xiaojie. Palladium Price Predic- tions via Machine Learning. Materials Circular Econ- omy, 6:32, 2024

2024
[27]

Machine Learning the Magnetocaloric Effect in Manganites from Lattice Pa- rameters

Zhang, Yun and Xu, Xiaojie. Machine Learning the Magnetocaloric Effect in Manganites from Lattice Pa- rameters. Applied Physics A , 126(4): , 2020

2020

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

S ¸enol Balaban, C

M. S ¸enol Balaban, C. Do˘ gulu, N. Akdede, H. Ako˘ glu, O. Karakayalı, S. Yılmaz, S. Yılmaz, T. Ajobiewe, S. G¨ uzel, G.˙Ikizer, et al. , “Emergency response, and community impact after february 6, 2023 kahraman- mara¸ s pazarcık and elbistan earthquakes: reconnais- sance findings and observations on affected region in t¨ urkiye,”Bulletin of Earthquake E...

work page 2023

[2] [2]

Heterogeneous multi-robot collaboration for coverage path planning in partially known dynamic environments,

G. G. R. de Castro, T. M. B. Santos, F. A. A. An- drade, J. Lima, D. B. Haddad, L. d. M. Hon´ orio, and M. F. Pinto, “Heterogeneous multi-robot collaboration for coverage path planning in partially known dynamic environments,” Machines, vol. 12, no. 3, 2024

work page 2024

[3] [3]

Cooperative aerial-ground multi- robot system for automated construction tasks,

M. Krizmancic, B. Arbanas, T. Petrovic, F. Petric, and S. Bogdan, “Cooperative aerial-ground multi- robot system for automated construction tasks,” IEEE Robotics and Automation Letters , vol. 5, no. 2, pp. 798–805, 2020

work page 2020

[4] [4]

Cooperative control of a heterogeneous multi- robot system based on relative localization,

M. Cognetti, G. Oriolo, P. Peliti, L. Rosa, and P. Ste- gagno, “Cooperative control of a heterogeneous multi- robot system based on relative localization,” in 2014 September 12, 2025 7:44 ws-us 14 REFERENCES IEEE/RSJ International Conference on Intelligent Robots and Systems , pp. 350–356, 2014

work page 2014

[5] [5]

Collaborative multi-robot search and rescue: Planning, coordination, perception, and active vision,

J. P. Queralta, J. Taipalmaa, B. Can Pullinen, V. K. Sarker, T. Nguyen Gia, H. Tenhunen, M. Gabbouj, J. Raitoharju, and T. Westerlund, “Collaborative multi-robot search and rescue: Planning, coordination, perception, and active vision,” IEEE Access, vol. 8, pp. 191617–191643, 2020

work page 2020

[6] [6]

Task assignment of multi- robot systems based on improved genetic algorithms,

S. Li, X. Xu, and L. Zuo, “Task assignment of multi- robot systems based on improved genetic algorithms,” in 2015 IEEE International Conference on Mechatron- ics and Automation (ICMA) , pp. 1430–1435, 2015

work page 2015

[7] [7]

A pso-based multi-robot cooperation method for target searching in unknown environments,

M. Dadgar, S. Jafari, and A. Hamzeh, “A pso-based multi-robot cooperation method for target searching in unknown environments,” Neurocomputing, vol. 177, pp. 62–74, 2016

2016

[8] [8]

Bal- ancing task allocation in multi-robot systems using k-means clustering and auction based mechanisms,

M. Elango, S. Nachiappan, and M. K. Tiwari, “Bal- ancing task allocation in multi-robot systems using k-means clustering and auction based mechanisms,” Expert Systems with Applications , vol. 38, no. 6, pp. 6486–6491, 2011

work page 2011