A Multi-Branch DQN-Based Transponder Resource Allocation Approach for Satellite Communications

Round 1

Reviewer 1 Report

Please see the attched file.

Comments for author File: Comments.pdf

Author Response

Please see the attachment.

Author Response File: Author Response.pdf

Reviewer 2 Report

The authors of this paper propose a reinforcement learning based Multi-Branch Deep Q-Network (MBDQN), which introduces TL-Branch and RP-Branch to extract features of satellite resource pool state and task state simultaneously, and Value-Branch to calculate the action-value function. The authors show that MBDQN improves the average resource occupancy performance (AOP) through the selection of multiple actions, including task selection and resource priority actions. In addition, the authors show that the trained MBDQN is more suitable for online deployment and significantly reduces the runtime overhead due to the fact that MBDQN does not need iteration in the test phase.

The subject of the paper is quite interesting and very well described and presented. Based on this, the reviewer believes that the paper can be accepted for publication.

During the preparation of the camera ready version, the authors should carefully revise the paper in order to correct some minor syntax/grammar errors, such as the one given below

Lines 185-186: Revise the sentence: “When a new resource leasing is required, we generate and format the… can be expressed as”.

Author Response

Please see the attachment.

Author Response File: Author Response.pdf