DeConNet: Deep Neural Network Model to Solve the Multi-Job Assignment Problem in the Multi-Agent System

Round 1

Reviewer 1 Report

  A deep neural network was proposed to solve the job assignment prob-321 lem between multiple agents and multiple jobs for a multi-agent system within a constant time. The execution time of the proposed DNN method was constant (ap-330 proximately 20 ms), whereas that of the algorithm used for the PDDL solver or the metaheuristic method in other studies increased rapidly according to the number of agents and jobs, and other parameters. This is an encouraging improvement so I suggest to accept it.

  But there are some things such as:

  1) The Dense-Concatenate Network (DeConNet) is depicted too rough so as to understand difficult.

  2) The planning domain definition language (PDDL) is illustrated too simple to be understood.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 2 Report

1)    The paper is grammatically poor and mostly inappropriate words are used.
•    In the abstract “was” is used for proposed solution and results. “In this study, a deep neural network (DNN) model 14 was proposed to solve the job assignment problem at a constant time regardless of the state of the 15 parameters”.
•    In the last paragraph of Introduction: word “report” is used instead of article or paper.
•    Related work. “In this section, we briefly survey related studies, including VRP variants, PDDL, and 80 machine learning for the assignment problem.
2)    Conclusion also not written well.
3)    Mostly references are older. So, add references of year 2019, 2020, 2021 and 2022 and add citations from reputed venues such as:

• A Compromise Programming to Task Assignment Problem in Software Development Project
• Metaheuristic Algorithms Based on Compromise Programming for the Multi-Objective Urban Shipment Problem

4)   Authors should include the units of values in the table either mention in the caption or table footer

5) Authors should highlight their contributions

6) The comparative analysis portion is weak. Authors should justify through proper justification about their results.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 3 Report

This research proposed a deep neural network, DeConNet, to solve the optimization problem with multi-job and multi-agent. This article is very interesting to read.

In results, the average travel time increased by a maximum of 13% compared to the ground cost. With the optimization strategy, the travel time was expected the reduce to a minimum value. Authors should provide more information and explanation to show the reason of the travel-time increasing with the optimization.

In line 138, “the execution time was reduced from tens of seconds to several minutes” is not reasonable.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 4 Report

The paper is well organised and clear, although the novelty of the method could be expressed more. There are several methods in the history for the same or similar problems like the "Hungarian" method by W. H. Kuhn (1955), inspired by Hungarian mathematics (Kőnig and Egerváry), or the well known method by W. R. Vogel (1958). The presented method is new, but could be explained why is it better.

Author Response

Please see the attachment.

Author Response File: Author Response.docx

Reviewer 5 Report

No comments

Author Response

Thank you for your assessment of the manuscript.

Round 2

Reviewer 3 Report

This research proposed a deep neural network to solve the optimization problem. This article is very interesting to read.