A biased random key genetic algorithm applied to the electric distribution network reconfiguration problem

This work presents a biased random-key genetic algorithm (BRKGA) to solve the electric distribution network reconfiguration problem (DNR). The DNR is one of the most studied combinatorial optimization problems in power system analysis. Given a set of switches of an electric network that can be opened or closed, the objective is to select the best configuration of the switches to optimize a given network objective while at the same time satisfying a set of operational constraints. The good performance of BRKGAs on many combinatorial optimization problems and the fact that it has never been applied to solve DNR problems are the main motivation for this research. A BRKGA is a variant of random-key genetic algorithms, where one of the parents used for mating is biased to be of higher fitness than the other parent. Solutions are encoded by using random keys, which are represented as vectors of real numbers in the interval (0,1), thus enabling an indirect search of the solution inside a proprietary search space. The genetic operators do not need to be modified to generate only feasible solutions, which is an exclusive task of the decoder of the problem. Tests were performed on standard distribution systems used in DNR studies found in the technical literature and the performance and robustness of the BRKGA were compared with other GA implementations.     

Biased random-key genetic algorithms for combinatorial optimization

Random-key genetic algorithms were introduced by Bean (ORSA J. Comput. 6:154–160, 1994) for solving sequencing problems in combinatorial optimization. Since then, they have been extended to handle a wide class of combinatorial optimization problems. This paper presents a tutorial on the implementation and use of biased random-key genetic algorithms for solving combinatorial optimization problems. Biased random-key genetic algorithms are a variant of random-key genetic algorithms, where one of the parents used for mating is biased to be of higher fitness than the other parent. After introducing the basics of biased random-key genetic algorithms, the paper discusses in some detail implementation issues, illustrating the ease in which sequential and parallel heuristics based on biased random-key genetic algorithms can be developed. A survey of applications that have recently appeared in the literature is also given.     

A biased random-key genetic algorithm for road congestion minimization

One of the main goals in transportation planning is to achieve solutions for two classical problems, the traffic assignment and toll pricing problems. The traffic assignment problem aims to minimize total travel delay among all travelers. Based on data derived from the first problem, the toll pricing problem determines the set of tolls and corresponding tariffs that would collectively benefit all travelers and would lead to a user equilibrium solution. Obtaining high-quality solutions for this framework is a challenge for large networks. In this paper, we propose an approach to solve the two problems jointly, making use of a biased random-key genetic algorithm for the optimization of transportation network performance by strategically allocating tolls on some of the links of the road network. Since a transportation network may have thousands of intersections and hundreds of road segments, our algorithm takes advantage of mechanisms for speeding up shortest-path algorithms.     

A biased random-key genetic algorithm with forward-backward improvement for the resource constrained project scheduling problem

This paper presents a biased random-key genetic algorithm for the resource constrained project scheduling problem. The chromosome representation of the problem is based on random keys. Active schedules are constructed using a priority-rule heuristic in which the priorities of the activities are defined by the genetic algorithm. A forward-backward improvement procedure is applied to all solutions. The chromosomes supplied by the genetic algorithm are adjusted to reflect the solutions obtained by the improvement procedure. The heuristic is tested on a set of standard problems taken from the literature and compared with other approaches. The computational results validate the effectiveness of the proposed algorithm.     

Finding multiple roots of a box-constrained system of nonlinear equations with a biased random-key genetic algorithm

Several numerical methods for solving nonlinear systems of equations assume that derivative information is available. Furthermore, these approaches usually do not consider the problem of finding all solutions to a nonlinear system. Rather, most methods output a single solution. In this paper, we address the problem of finding all roots of a system of equations. Our method makes use of a biased random-key genetic algorithm (BRKGA). Given a nonlinear system, we construct a corresponding optimization problem, which we solve multiple times, making use of a BRKGA, with areas of repulsion around roots that have already been found. The heuristic makes no use of derivative information. We illustrate the approach on seven nonlinear equations systems with multiple roots from the literature.     

A biased random-key genetic algorithm for the Steiner triple covering problem

We present a biased random-key genetic algorithm (BRKGA) for finding small covers of computationally difficult set covering problems that arise in computing the 1-width of incidence matrices of Steiner triple systems. Using a parallel implementation of the BRKGA, we compute improved covers for the two largest instances in a standard set of test problems used to evaluate solution procedures for this problem. The new covers for instances A ₄₀₅ and A ₇₂₉ have sizes 335 and 617, respectively. On all other smaller instances our algorithm consistently produces covers of optimal size.     

The use of random weights for the training of multilayer networks of neurons with Heaviside characteristics

Artificial neural networks have, in recent years, been very successfully applied in a wide range of areas. A major reason for this success has been the existence of a training algorithm called backpropagation. This algorithm relies upon the neural units in a network having input/output characteristics that are continuously differentiable. Such units are significantly less easy to implement in silicon than are neural units with Heaviside (step-function) characteristics. In this paper, we show how a training algorithm similar to backpropagation can be developed for 2-layer networks of Heaviside units by treating the network weights (i.e., interconnection strengths) as random variables. This is then used as a basis for the development of a training algorithm for networks with any number of layers by drawing upon the idea of internal representations. Some examples are given to illustrate the performance of these learning algorithms.     

A biased random-key genetic algorithm for the project scheduling problem with flexible resources

In this paper, we investigate a resource-constrained project scheduling problem with flexible resources. This is an \(\mathcal {NP}\)-hard combinatorial optimization problem that consists of scheduling a set of activities requiring specific resource units of several skills. The goal is to minimize the makespan of the project. We propose a biased random-key genetic algorithm for computing feasible solutions for the referred problem. We study different decoding mechanisms: an already existing method in the literature, a new adapted serial scheduling generation scheme, and a combination of both. The new procedure is tested using a set of benchmark instances of the problem. The results provide strong evidence that the new heuristic is robust and yields high-quality feasible solutions.     

Maximizing Queueing Network Utility Subject to Stability: Greedy Primal-Dual Algorithm

We study a model of controlled queueing network, which operates and makes control decisions in discrete time. An underlying random network mode determines the set of available controls in each time slot. Each control decision “produces” a certain vector of “commodities”; it also has associated “traditional” queueing control effect, i.e., it determines traffic (customer) arrival rates, service rates at the nodes, and random routing of processed customers among the nodes. The problem is to find a dynamic control strategy which maximizes a concave utility function H(X), where X is the average value of commodity vector, subject to the constraint that network queues remain stable.We introduce a dynamic control algorithm, which we call Greedy Primal-Dual (GPD) algorithm, and prove its asymptotic optimality. We show that our network model and GPD algorithm accommodate a wide range of applications. As one example, we consider the problem of congestion control of networks where both traffic sources and network processing nodes may be randomly time-varying and interdependent. We also discuss a variety of resource allocation problems in wireless networks, which in particular involve average power consumption constraints and/or optimization, as well as traffic rate constraints.     

Randomized heuristics for handover minimization in mobility networks

A mobile device connects to the cell tower (base station) from which it receives the strongest signal. As the device moves it may connect to a series of towers. The process in which the device changes the base station it is connected to is called handover. A cell tower is connected to a radio network controller (RNC) which controls many of its operations, including handover. Each cell tower handles an amount of traffic and each radio network controller has capacity to handle a maximum amount of traffic from all base stations connected to it. Handovers between base stations connected to different RNCs tend to fail more often than handovers between base stations connected to the same RNC. Handover failures result in dropped connections and therefore should be minimized. The Handover Minimization Problem is to assign towers to RNCs such that RNC capacity is not violated and the number of handovers between base stations connected to different RNCs is minimized. We describe an integer programming formulation for the handover minimization problem and show that state-of-the-art integer programming solvers can solve only very small instances of the problem. We propose several randomized heuristics for finding approximate solutions of this problem, including a GRASP with path-relinking for the generalized quadratic assignment problem, a GRASP with evolutionary path-relinking, and a biased random-key genetic algorithm. Computational results are presented.     

Flooding-limited and multi-constrained QoS multicast routing based on the genetic algorithm for MANETs

A wireless MANET is a collection of wireless mobile hosts that dynamically create a temporary network without a fixed infrastructure. The topology of the network may change unpredictably and frequently. Therefore, multicast routing in ad hoc networks is a very challenging problem. This paper proposes a multi-constrained QoS multicast routing method using the genetic algorithm. The proposal will be flooding-limited using the available resources and minimum computation time in a dynamic environment. By selecting the appropriate values for parameters such as crossover, mutation, and population size, the genetic algorithm improves and tries to optimize the routes. Simulation results indicate its better performances compared to other methods.     

A Neural Network Application in Personnel Scheduling

In this paper, we present and evaluate a neural network model for solving a typical personnel-scheduling problem, i.e. an airport ground staff rostering problem. Personnel scheduling problems are widely found in servicing and manufacturing industries. The inherent complexity of personnel scheduling problems has normally resulted in the development of integer programming-based models and various heuristic solution procedures. The neural network approach has been admitted as a promising alternative to solving a variety of combinatorial optimization problems. While few works relate neural network to applications of personnel scheduling problems, there is great theoretical and practical value in exploring the potential of this area. In this paper, we introduce a neural network model following a relatively new modeling approach to solve a real rostering case. We show how to convert a mixed integer programming formulation to a neural network model. We also provide the experiment results comparing the neural network method with three popular heuristics, i.e. simulated annealing, Tabu search and genetic algorithm. The computational study reveals some potential of neural networks in solving personnel scheduling problems.     

Heuristic approaches to determine base-stock levels in a serial supply chain with a single objective and with multiple objectives

This paper deals with the problem of determination of installation base-stock levels in a serial supply chain. The problem is treated first as a single-objective inventory-cost optimization problem, and subsequently as a multi-objective optimization problem by considering two cost components, namely, holding costs and shortage costs. Variants of genetic algorithms are proposed to determine the best base-stock levels in the single-objective case. All variants, especially random-key gene-wise genetic algorithm (RKGGA), show an excellent performance, in terms of convergence to the best base-stock levels across a variety of supply chain settings, with minimum computational effort. Heuristics to obtain base-stock levels are proposed, and heuristic solutions are introduced in the initial population of the RKGGA to expedite the convergence of the genetic search process. To deal with the multi-objective supply-chain inventory optimization problem, a simple multi-objective genetic algorithm is proposed to obtain a set of non-dominated solutions.     

Automatic structure and parameter training methods for modeling of mechanical systems by recurrent neural networks

Automatic nonlinear-system identification is very useful for various disciplines including, e.g., automatic control, mechanical diagnostics and financial market prediction. This paper describes a fully automatic structural and weight learning method for recurrent neural networks (RNN). The basic idea is training with residuals, i.e., a single hidden neuron RNN is trained to track the residuals of an existing network before it is augmented to the existing network to form a larger and, hopefully, better network. The network continues to grow until either a desired level of accuracy or a preset maximal number of neurons is reached. The method requires no guessing of initial weight values or the number of neurons in the hidden layer from users. This new structural and weight learning algorithm is used to find RNN models for a two-degree-of-freedom planar robot, a Van der Pol oscillator and a Mackey–Glass equation using their simulated responses to excitations. The algorithm is able to find good RNN models in all three cases.     

A memetic algorithm for the virtual network mapping problem

The Internet has ossified. It has lost its capability to adapt as requirements change. A promising technique to solve this problem is the introduction of network virtualization. Instead of directly using a single physical network, working just well enough for a limited range of applications, multiple virtual networks are embedded on demand into the physical network, each of them perfectly adapted to a specific application class. The challenge lies in mapping the different virtual networks with all the resources they require into the available physical network, which is the core of the virtual network mapping problem. In this work, we introduce a memetic algorithm that significantly outperforms the previously best algorithms for this problem. We also offer an analysis of the influence of different problem representations and in particular the implementation of a uniform crossover for the grouping genetic algorithm that may also be interesting outside of the virtual network mapping domain. Furthermore, we study the influence of different hybridization techniques and the behaviour of the developed algorithm in an online setting.     

Ranking hubs and authorities using matrix functions

The notions of subgraph centrality and communicability, based on the exponential of the adjacency matrix of the underlying graph, have been effectively used in the analysis of undirected networks. In this paper we propose an extension of these measures to directed networks, and we apply them to the problem of ranking hubs and authorities. The extension is achieved by bipartization, i.e., the directed network is mapped onto a bipartite undirected network with twice as many nodes in order to obtain a network with a symmetric adjacency matrix. We explicitly determine the exponential of this adjacency matrix in terms of the adjacency matrix of the original, directed network, and we give an interpretation of centrality and communicability in this new context, leading to a technique for ranking hubs and authorities. The matrix exponential method for computing hubs and authorities is compared to the well known HITS algorithm, both on small artificial examples and on more realistic real-world networks. A few other ranking algorithms are also discussed and compared with our technique. The use of Gaussian quadrature rules for calculating hub and authority scores is discussed.     

A Zoom-In Approach to Design SDH Mesh Restorable Networks

Mesh restorable networks based on SONET (Synchronous Optical Network, standard optical transmission technology widely accepted and implemented in North America) or SDH (Synchronous Digital Hierarchy, the European standard currently adopted by the major European telecom operators) are an economically attractive solution in areas where high demand and high connectivity are involved (Wu, 1995). In these networks, the reconfiguration capability of the digital cross connect systems (DCS) allows to reroute the demand affected by network failures. The degree of sharing of spare capacity in networks based on this architecture is high.This paper presents a heuristic algorithm for solving the near-optimal design of SDH mesh-type link restorable networks, i.e. determining the network topology and assigning the capacity to transport the demand in normal situations and to allow full link restorability in case of single link failures. The algorithm is based on a Zoom-In technique, a novel approach which forms a compromise between sequential and integrated techniques. The different building blocks of the algorithm are tested extensively and compared with other results mentioned in literature. Comparison of the simulation results for the overall design problem with other solution techniques indicates that the Zoom-In method is a quite promising approach, able to combine the accuracy of integrated approaches with the calculation speed of sequential approaches.     

Farsighted free trade networks

The paper examines whether bilateral free trade agreements can lead to global free trade. We reconsider the endogenous tariff model introduced by Goyal and Joshi (2006) who study pairwise stability of free trade networks. We depart from their analysis by adopting the concept of pairwise farsightedly stable networks (Herings et al. 2009, GEB). We show that the complete network (i.e., global free trade) constitutes a pairwise farsightedly stable set. In particular, there is a farsightedly improving path from the empty network (i.e., no free trade agreement in place) to the complete network, which involves link additions only, while farsightedly improving paths from preexisting free trade networks may involve link deletion (i.e., dissolution of some bilateral FTAs). Moreover, we show that pairwise farsightedly stable set of networks is not unique. One implication of our results is that bilateral trade negotiations, if properly channeled, can lead to global free trade, although some bilateral agreements may have to be dissolved first to pave the way towards global free trade.     

Analysis of Steiner subtrees of random trees for traceroute algorithms

We consider in this article the problem of discovering, via a traceroute algorithm, the topology of a network, whose graph is spanned by an infinite branching process. A subset of nodes is selected according to some criterion. As a measure of efficiency of the algorithm, the Steiner distance of the selected nodes, i.e. the size of the spanning subtree of these nodes, is investigated. For the selection of nodes, two criteria are considered: a node is randomly selected with a probability, which is either independent of the depth of the node (uniform model) or else in the depth biased model, is exponentially decaying with respect to its depth. The limiting behavior the size of the discovered subtree is investigated for both models. © 2009 Wiley Periodicals, Inc. Random Struct. Alg., 2009