Integrating Tabu Search in Particle Swarm Optimization for the Frequency Assignment Problem

In this paper, we address one of the issues in the frequency assignment problem for cellular mobile networks in which we intend to minimize the interference levels when assigning frequencies from a limited frequency spectrum. In order to satisfy the increasing demand in such cellular mobile networks, we use a hybrid approach consisting of a Particle Swarm Optimization(PSO) combined with a Tabu Search(TS) algorithm. This approach takes both advantages of PSO efficiency in global optimization and TS in avoiding the premature convergence that would lead PSO to stagnate in a local minimum. Moreover, we propose a new efficient, simple, and inexpensive model for storing and evaluating solution's assignment. The purpose of this model reduces the solution's storage volume as well as the computations required to evaluate thesesolutions in comparison with the classical model. Our simulation results on the most known benchmarking instances prove the effectiveness of our proposed algorithm in comparison with previous related works in terms of convergence rate, the number of iterations, the solution storage volume and the running time required to converge to the optimal solution.

Adaptive genetic algorithms guided by decomposition for PCSPs： application to frequency assignment problems

This paper proposes Adaptive Genetic Algorithms Guided by structural knowledges coming from decomposition methods, for solving PCSPs. The family of algorithms called AGAGD_x_y is designed to be doubly genetic, meaning that any decomposition method and different heuristics for the genetic operators can be considered. To validate the approach, the decomposition algorithm due to Newman was used and several crossover operators based on structural knowledge such as the cluster, separator and the cut were tested. The experimental results obtained on the most challenging Minimum Interference-FAP problems of CALMA instances are very promising and lead to interesting perspectives to be explored in the future.

Real edge spans of distance two labelings of graphs

An L（j, k）-labeling of a graph G is an assignment of nonnegative integers to the vertices of G such that adjacent vertices receive integers which are at least j apart, and vertices at distance two receive integers which are at least k apart. Given an L（j, k）-labeling f of G, define the L（j, k） edge span of f, βj,k（G,f） =max{ ｜f（x）-f（y）｜： {x,y}∈E（G）}. The L（j,k） edge span of G, βj,k （G） is min βj,k（ G, f）, where the minimum runs over all L（j, k）-labelings f of G. The real L（.j, k）-labeling of a graph G is a generalization of the L（j, k）-labeling. It is an assignment of nonnegative real numbers to the vertices of G satisfying the same distance one and distance two conditions. The real L（j, k） edge span of a graph G is defined accordingly, and is denoted by βj,k（G）. This paper investigates some properties of the L（j, k） edge span and the real L（j, k） edge span of graphs, and completely determines the edge spans of cycles and complete t-partite graphs.