Search

Policy gradient algorithms suffer from the large variance of gradient estimation. the algorithm of policy gradient with parameter based exploration mitigates this problem to some extent. However, PGPE estimates its gradient based on the Monte Carlo, which requires a large number of samples to achieve the fairly stable policy update. And thus hinders its application in the real world problem. In order to further reduce the variance of policy gradient, the algorithm of function approximation for policy gradients with parameter-based exploration (PGPE-FA) implements the algorithm of PGPE in the Actor-Critic framework. More specifically, the proposed method utilized value function to estimate the policy gradient, instead of using trajectory samples to estimate the policy gradient as PGPE method does, thereby reducing the variance of gradient estimation. Finally, the experiment verifies that the proposed algorithm can reduce the variance of gradient estimation. [ABSTRACT FROM AUTHOR]

Copyright of Operations Research Transactions / Yunchouxue Xuebao is the property of Editorial office of Operations Research Transactions and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Systems Engineering & Electronics is the property of Journal of Systems Engineering & Electronics Editorial Department and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

With the increasing maturity of wireless charging technology, especially the development of magnetic resonance wireless charging technology, the employment of mobile charging vehicles and wireless charging technology to replenish sensors' energy to ensure the continuous operation of wireless sensor networks has become a new research hotspot. This paper focused on how to schedule multiple charging vehicles to replenish energy to the sensors in a large-scale wireless sensor network. In order to balance the charging tasks of multiple charging vehicles and reduce the completion time of the whole charging task, the paper proposed the charging task completion time minimization problem, hoping to find K closed charging circles for K charging vehicles, so that the longest completion time among these K vehicles was the shortest. Since the charging task completion time minimization problem was an NP-hard problem and it was difficult to find an optimal solution in polynomial time, this paper proposed an approximation algorithm with a ratio of 5 for this problem. Finally, it demonstrated the performance of the algorithm by simulation experiments, and the experiments show that the actual approximation ratio of the proposed algorithm is less than 2. [ABSTRACT FROM AUTHOR]

Copyright of Operations Research Transactions / Yunchouxue Xuebao is the property of Editorial office of Operations Research Transactions and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

The virtual backbone (VB) in wireless networks is a subset of some wireless nodes, such that only VB nodes are responsible for routing related tasks. The smaller the weight of VB, the less the overhead. In a node-weighted wireless network, not only the number of nodes in VB should be considered, but also the total weight should be considered. Frequently, a weighted wireless network is modeled as a node-weighted unit disk graph ( UDG) . Correspondingly, the minimum-weighted VB problem in the weighted wireless network is abstracted as the minimum-weighted connected dominating set ( MWCDS) problem in the node-weighted UDG. Finding MWCDS is a NP-hard problem. To reduce the approximate ratio of the connected part of the MW CDS problem in node-weighted UDG, this paper proposed a new method-degree-based node-weighted Steiner tree algorithm in the connected part. Combining the best results at present, it will obtain a ( 3. 32 + s ) -approximation algorithm for MWCDS problem in UDG. Similarly, it will also obtain a ( 3. 3 2 + s) -approximation algorithm for MWCVC problem in UDG. It is proved that the method of reducing the approximate ratio of the MWCDS problem in node-weighted UDG is feasible by improving the approximate ratio of connected parts. [ABSTRACT FROM AUTHOR]

In today' s big data era, big data processing frameworks such as MapReduce often appear slow and inefficient when processing data, specially related to graphs. Therefore, it is necessary to explore an efficient algorithm to handle this type of clique counting problem. Since the predecessor literatures have thoroughly explored the 3-clique counting, the extended version of the problem( the 4-clique counting problem) improves its position gradually. Under the guidance of a heuristic idea, this paper proposed a probability sampling algorithm based on neighboring edge sampling to solve the extended problem. With the usage of Chernoff inequality, the algorithm only needed a certain number of samplers as the performance guarantee of relative error under the approximate condition. Later, the experimental evaluation and comparison shows that the probability sampling algorithm loses a small amount of precision compared with the traditional precision algorithm, but it has great advantages in algorithm running time and space occupation. Finally, it comes to the conclusion that it has great practical value in practical applications. [ABSTRACT FROM AUTHOR]

A virtual backbone of wireless sensor networks is a subset of nodes and only virtual backbone nodes are responsible for routing-related tasks. The smaller a virtual backbone is, the less the network overhead, the size of a virtual backbone is the key factor to measure the quality of the vi11ual backbone. Frequently, the unit disk graph is used to model a wireless sensor network. In a wireless sensor network, the problem of finding the minimum virtual backbone is abstracted to the problem of finding the minimum connected dominating set in the unit disk graph corresponding to the wireless network. However, the problem of finding the minimum connected dominating set in the unit disk graph is a NP-hard problem, many works contribute to obtain the approximation algorithm of the minimum connected dominating set rather than finding the minimum connected dominating set in the unit disk graph. Constructing a 3-connected r-hop dominating set in wireless sensor networks is an effective means to obtain a connected dominating set with a smaller size such that can reduce the total amount of information forwarded between any nodes, which is a better approximation for finding the minimum virtual backbone. This paper proposed an algorithm to construct a 3-connected r-hop dominating set in wireless sensor network. It shows that the size of a 3-connected r-hop dominating set produced by the algorithm is no more than 5 (2r + 2β + 1) (r + 1) β| U' I - 10(2 + β) (r + 1) - 5r - 12. At the end, it analyzed the performance of the proposed algorithm by experiment simulation, and the experimental results are consistent with what expected. [ABSTRACT FROM AUTHOR]

The successive cancellation(SC) decoding method for polar codes under additive white Gaussian noise (AWGN) channels is performed in the log-likelihood ratio(LLR) domain. The calculation of the f function nodes use a sum-product algorithm based on hyperbolic tangent rules. This paper proposed a polyline approximation algorithm, which simplified the hyperbolic tangent function and the inverse hyperbolic tangent function into a 9-segment polyline function respectively. In order to obtain better FER performance under the polyline approximation algorithm, this algorithm added a 16 bit CRC to the information bits before encoding. Simulation experiments show that for the polar codes with code length N = 1 024 and information bit length K = 496, the proposed algorithm has better FER performance than the sum-product algorithm, and it reduces the decoding complexity and improves the decoding speed. [ABSTRACT FROM AUTHOR]

Copyright of Electric Machines & Control / Dianji Yu Kongzhi Xuebao is the property of Electric Machines & Control and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Based on the classical problem of project portfolio selection (PPS), the multi-phase rolling model of PPS was constructed under uncertainty, the objective was maximization of benefits of each phase. Related constraints were built relying both matching of strategy equilibrium between enterprise strategy requirement and synthesis of strategy contribution of projects which were selected, and the enhancement effect of resources. Upper-side and down-side nearness were put up based on fuzzy theory, and was used to measured the matching of strategy equilibrium. A crisp transformation was implemented with the fuzzy model based on possibility theory. Genetic algorithm was presented to solved the model, through development of path relinking (PR), the algorithm of successive substitute and successive retroposition was designed to applied to 0-1 encoding as local search, and get the better solution. Simulation test show that when project portfolio selection of multi phase rolling is doing, enterprise can get benefits steadily during five phases, and benefits is differentiate in 6-8 phases. This conclusion can help enterprise adjust its strategy periods. Statistics about numbers of project styles are selected in every phase show the periodic regularity. Research resultion can give a guidance to execution of project portfolio selection in fact. [ABSTRACT FROM AUTHOR]

To obtain high quality finite difference mesh for castings with complex geometric shapes and thin walls, and to establish an effective algorithm for finite difference mesh generation of STL solid models, an enmeshment process of multiple STL models based on slice line scanning method was described and a computer program for finite difference mesh generation was developed. The program adopted dynamic data structure, updated technique of treating singular points and mesh filtering method for fault tolerance. It can be applied to generate 3D non-uniform meshes for assembled STL solid models rapidly. The algorithm is fast in running, alterable in mesh spacing, optimizational in singular points treating, and high fault tolerance. The mesh generation is rapid, easy to use, low memory needed, accurate and reliable. [ABSTRACT FROM AUTHOR]