OPTIMIZATION OF THE PLACEMENT OF WI-FI ACCESS POINTS TAKING INTO ACCOUNT THE ETHERNET TRUNK IN THE CORPORATE NETWORK

Abstract

The article addresses the applied problem of optimizing Wi-Fi access point placement in a corporate network while considering not only radio coverage but also spatial distribution of user demand and capacity of the Ethernet backbone. The relevance of the study is determined by the fact that traditional approaches to wireless infrastructure planning usually focus on signal strength or the number of active access points, whereas wired segment constraints, uplink capacity and uneven traffic demand are often excluded from a single evaluation model. The purpose of the work is to formalize and experimentally verify a model for selecting the optimal number and location of access points that can provide a balanced combination of coverage, throughput and load distribution. Within the study, a simulation model was developed in which the indoor environment is represented as a set of spatial cells, candidate access points are derived from Wi-Fi fingerprints of the UjiIndoorLoc dataset, and configuration quality is assessed through an integrated objective function. This function accounts for the share of covered zones, the share of served demand, Ethernet-backbone utilization, the number of active APs and the overload level. To evaluate the proposed approach, the algorithm was compared with random selection, a coverage-only greedy algorithm and strongest-signal-based selection. The simulation results showed that the integrated approach reaches a high objective-function value faster, provides a higher share of served demand and ensures more controlled use of the Ethernet backbone. This makes the model suitable for preliminary planning of corporate wireless infrastructure. Scientific novelty consists in combining Wi-Fi coverage, user load and backbone throughput criteria within one optimization model.