The paper presents a novel localization method called MDS-MAP, which uses connectivity information to estimate the geographic positions of nodes in a communication network. Unlike methods that require GPS receivers or other sophisticated sensors, MDS-MAP leverages multidimensional scaling (MDS), a data analysis technique, to derive node locations based on the connectivity information. The method can incorporate additional information such as estimated distances between neighbors or known positions of anchor nodes if available. The algorithm is efficient, with a time complexity of \(O(n^3)\) for a network of \(n\) nodes. Through simulations, the authors demonstrate that MDS-MAP is more robust to measurement errors compared to previous methods, especially when nodes are uniformly distributed. It also requires fewer anchor nodes to achieve comparable results and can provide relative coordinates even without anchor nodes. The paper discusses the method's performance on various network topologies and placement scenarios, showing that it performs well with low connectivity levels and small placement errors. The authors also explore possible extensions, including distributed implementation and the use of more advanced MDS techniques.The paper presents a novel localization method called MDS-MAP, which uses connectivity information to estimate the geographic positions of nodes in a communication network. Unlike methods that require GPS receivers or other sophisticated sensors, MDS-MAP leverages multidimensional scaling (MDS), a data analysis technique, to derive node locations based on the connectivity information. The method can incorporate additional information such as estimated distances between neighbors or known positions of anchor nodes if available. The algorithm is efficient, with a time complexity of \(O(n^3)\) for a network of \(n\) nodes. Through simulations, the authors demonstrate that MDS-MAP is more robust to measurement errors compared to previous methods, especially when nodes are uniformly distributed. It also requires fewer anchor nodes to achieve comparable results and can provide relative coordinates even without anchor nodes. The paper discusses the method's performance on various network topologies and placement scenarios, showing that it performs well with low connectivity levels and small placement errors. The authors also explore possible extensions, including distributed implementation and the use of more advanced MDS techniques.