.png){kind=icon}
.png){kind=icon}
Return
Modeling 5G shared base station planning problem using an evolutionary bi-level optimization algorithm☆ ☆
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2024.112079.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
With the cost of 5G network construction surges, Base Station (BS) sharing is becoming more and more popular among operators nowadays. A typical scenario of 5G shared BS planning is presented in this paper, in which different operators share the BSs constructed by the same tower company to reduce the investments. As a result, the 5G shared BS planning problem is modeled as a Bi-Level Optimization Problem (BLOP), with the tower company as the upper level decision-maker and the operators as the lower level decision-makers. The suggested bi-level 5G shared BS planning model considers constraints of various kinds involved in the real network planning scenario, therefore making it more realistic and applicable. Considering the impact of constraints, a Transfer Learning based Evolutionary Algorithm for shared BS Planning (TLEA-BSP) is introduced to solve the proposed bi-level 5G shared BS planning model. In TLEA-BSP, infeasible and feasible information can be transferred and utilized among the optimization processes of multiple lower level problems, such that the upper level and lower level optimization efficiency can be greatly improved. Experimental simulations are conducted on ten generated instances with different scales, and the effectiveness of the proposed model is confirmed by comparing the results of the bi-level model with those of two single-level 5G BS planning models. Comparing the proposed algorithm with the well-established BL-CMA-ES without using the transfer learning strategy for the bi-level 5G shared BS planning model also reveals the essential role of the transfer learning strategy.
Keywords:
Bi-level optimization
Evolutionary algorithm
Modeling
Base station sharing
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W