Pareto Optimal Network Research Articles

Multiobjective bilevel programming model for multilayer perceptron neural networks

The architecture of multilayer perceptron (MLP) neural networks dictates the network's performance. However, aiming at the specific classification problems, suitable architectures of MLPs must be determined beforehand. More often, the design decisions rely on a trial and error learning and the experience knowledge. To automatically design an MLP's network architecture and optimize network parameters, a multiobjective bilevel programming model is built. In this model, a multiobjective optimization problem is constructed in the upper level for obtaining a set of Pareto optimal architectures of the MLPs, considering network complexity, training error rate, and validation error rate, while a single-objective optimization problem is established in the lower level to search for the optimum network parameters for a given network architecture. For dealing with this model efficiently, a novel multiobjective hierarchical learning algorithm (MOHLA) is proposed, in which an integer-coding NSGA-II is developed as the upper-level optimizer for a set of Pareto optimal network structures of the MLPs, while a non-iterative method is regarded as the lower-level solver for the MLPs' connection parameters. After a set of trained MLPs is obtained finally by using MOHLA, a selective ensemble strategy is adopted for improving identification accuracy. Three types of multiobjective bilevel programming models are investigated and compared in the experiments. Moreover, the proposed MOHLA is compared with several state-of-the-art learning approaches on various classification problems. The experimental results confirm that MOHLA performs well.

Multiobjective Brain Storm Optimization Community Detection Method Based on Novelty Search

A complex network is characterized by community structure, so it is of great theoretical and practical significance to discover hidden functions by detecting the community structure in complex networks. In this paper, a multiobjective brain storm optimization based on novelty search (MOBSO-NS) community detection method is proposed to solve the current issue of premature convergence caused by the loss of diversity in complex network community detection based on multiobjective optimization algorithm and improve the accuracy of community discovery. The proposed method designs a novel search strategy where novelty individuals are first constructed to improve the global search ability, thus avoiding falling into local optimal solutions; then, the objective space is divided into 3 clusters: elite cluster, ordinary cluster, and novel cluster, which are mapped to the decision space, and finally, the populations are disrupted and merged. In addition, the introduction of a restarting strategy is introduced to avoid stagnation by premature convergence. Experimental results show that the algorithm with good global searchability can find the Pareto optimal network community structure set with uniform distribution and high convergence and excavate the network community with higher quality.

Multi-objective optimization framework complying IEEE 802.15.6 communication standards for wireless body area networks

Wireless body area network (WBAN) routing protocols are primarily designed for improvement of network performance parameters such as network lifetime, throughput and latency. However, these designs do not take into account the health hazards posed due to WBAN radiation. The focus of implementing IEEE 802.15.6 standards is to develop an energy-efficient, reliable and health-conscious wireless communication system around human body. This paper presents the implementation of a clustering-based routing protocol for WBAN that targets in achieving overall optimization of WBAN performance parameters including network lifetime, throughput, latency, node signal power and specific absorption rate (SAR) of human body for emf radiation. The suggested protocol applies a multi-objective NSGA-II optimization heuristic for cluster formation while taking into account various IEEE 802.15.6 standard constraints. A unique chromosome structure has been proposed for NSGA-II population in which each chromosome represents a random network topology for determining the node clusters, cluster heads and node transmission power. Furthermore, three system-level models for conducting multi-objective fitness evaluation of each chromosome have been proposed. These models predict the global transmission energy consumption; the average received signal strength (RSSI) and end to end network latency respectively. The developed NSGA-II model evolves into a non-dominated set of Pareto-optimal network topologies that renders the desired objectives of minimization of network energy consumption and network latency besides maximization of average RSSI. Enhancement in RSSI further leads to a tremendous improvement in the throughput rate. The node transmission power specifications of Pareto-optimal network topologies meet the standard SAR constraints too. The performance results shows that the suggested protocol strictly meets IEEE 802.15.6 compliance.

An Efficient Multi-Objective Optimization Method for Use in the Design of Marine Protected Area Networks

An efficient connectivity-based method for multi-objective optimisation applicable to the design of marine protected area networks is described. Multi-objective network optimisation highlighted previously unreported step changes in the structure of optimal subnetworks for protection associated with minimal changes in cost or benefit functions. This emphasises the desirability of performing a full, unconstrained, multi-objective optimisation for marine spatial planning. Brute force methods, examining all possible combinations of protected and unprotected sites for a network of sites, are impractical for all but the smallest networks as the number of possible networks grows as 2^m, where m is the number of sites within the network. A metaheuristic method based around Markov Chain Monte Carlo methods is described which searches for the set of Pareto optimal networks (or a good approximation thereto) given two separate objective functions, for example for network quality or effectiveness, population persistence, or cost of protection. The optimisation and search methods are independent of the choice of objective functions and can be easily extended to more than two functions. The speed, accuracy and convergence of the method under a range of network configurations are tested with model networks based on an extension of random geometric graphs. Examination of two real-world marine networks, one designated for the protection of the stony coral Lophelia pertusa, the other a hypothetical man-made network of oil and gas installations to protect hard substrate ecosystems, demonstrates the power of the method in finding multi-objective optimal solutions for networks of up to 100 sites. Results using network average shortest path as a proxy for population resilience and gene flow within the network supports the use of a conservation strategy based around highly connected clusters of sites.

Application of SNODAS and hydrologic models to enhance entropy-based snow monitoring network design

Snow has a unique characteristic in the water cycle, that is, snow falls during the entire winter season, but the discharge from snowmelt is typically delayed until the melting period and occurs in a relatively short period. Therefore, reliable observations from an optimal snow monitoring network are necessary for an efficient management of snowmelt water for flood prevention and hydropower generation. The Dual Entropy and Multiobjective Optimization is applied to design snow monitoring networks in La Grande River Basin in Québec and Columbia River Basin in British Columbia. While the networks are optimized to have the maximum amount of information with minimum redundancy based on entropy concepts, this study extends the traditional entropy applications to the hydrometric network design by introducing several improvements. First, several data quantization cases and their effects on the snow network design problems were explored. Second, the applicability the Snow Data Assimilation System (SNODAS) products as synthetic datasets of potential stations was demonstrated in the design of the snow monitoring network of the Columbia River Basin. Third, beyond finding the Pareto-optimal networks from the entropy with multi-objective optimization, the networks obtained for La Grande River Basin were further evaluated by applying three hydrologic models. The calibrated hydrologic models simulated discharges using the updated snow water equivalent data from the Pareto-optimal networks. Then, the model performances for high flows were compared to determine the best optimal network for enhanced spring runoff forecasting.

A Game Theoretical Approach for Topology Control in Wireless Ad Hoc Networks with Selfish Nodes

In ad hoc networks, a significant amount of energy available to devices is utilized in network management operations. Since devices have limited energy resources, therefore, they drop data packets of other nodes to reduce their energy consumption. This selfish behaviour increases number of retransmissions over the link which increases energy consumption of the source node, introduces time delays, and degrades throughput of the network. Although conventional distributed topology control solutions minimize energy utilization of the nodes by adjustment of transmission power, however, selfish behaviour by devices introduce additional complexity in design which make topology control a challenging task. In this paper, we proposed Energy Efficient Topology Control Algorithm (EETCA) using game theoretical approach, in which, utility of the node depends on selfishness of the neighbors, link traffic rate, and link length. In decision-making step, nodes remove the links with other nodes that have high drop rate under the condition that network remains connected. We show that Nash Equilibrium point of the proposed game results in Pareto optimal network topology. We compare results of EETCA with Optimum (OPT) and Minimum Least Power Path Tree (MLPT) algorithms presented in literature. We carried our simulations under multiple sources scenario which show that EETCA outperforms previous approaches when number of nodes in the network increases. Furthermore, we simulate the performance of Ad-hoc On-demand Distance Vector (AODV) routing protocol under EETCA topology and compare it with MLPT and OPT topologies. The results show that the ad hoc network constructed using proposed solution substantially improves throughput of AODV routing protocol as compared to MLPT and OPT topology control algorithms.

Limits and trade-offs of topological network robustness.

We investigate the trade-off between the robustness against random and targeted removal of nodes from a network. To this end we utilize the stochastic block model to study ensembles of infinitely large networks with arbitrary large-scale structures. We present results from numerical two-objective optimization simulations for networks with various fixed mean degree and number of blocks. The results provide strong evidence that three different blocks are sufficient to realize the best trade-off between the two measures of robustness, i.e. to obtain the complete front of Pareto-optimal networks. For all values of the mean degree, a characteristic three block structure emerges over large parts of the Pareto-optimal front. This structure can be often characterized as a core-periphery structure, composed of a group of core nodes with high degree connected among themselves and to a periphery of low-degree nodes, in addition to a third group of nodes which is disconnected from the periphery, and weakly connected to the core. Only at both extremes of the Pareto-optimal front, corresponding to maximal robustness against random and targeted node removal, a two-block core-periphery structure or a one-block fully random network are found, respectively.

Multi-Objective Optimization of Network Systems by Using ANT Algorithms

At the design stage, the availability of a network system can be improved by redundancy allocation or, for a fixed network topology, by selecting highly available elements as nodes and connection arcs. Yet these design choices are constrained by the available budget and other physical and logistic requirements. This paper formulates the network design problem as a multiple-objective optimization aiming at maximizing the network availability and minimizing its associated cost. The potential of using Ant algorithms to identify Pareto-optimal network designs with respect to the defined objectives is explored. For each solution, the network availability objectives is computed by a combination of Monte Carlo simulation and Cellular Automata. The Pareto-optimal solutions can be used by the decision-makers to identify compromise solutions which best satisfy their risk profiles. A sample network is solved as a demonstration of the proposed approach.

COORDINATION FAILURES IN NETWORK MIGRATION*

Previous immigration facilitates future immigration, a phenomenon called network migration. While well recognized, previous research has mainly focused on the implications of immigrant networks on future migrants. In contrast, this paper derives a simple model from the perspective of the incumbent immigrant population by introducing sub‐networks and argues that the incumbent migrants fail to achieve a Pareto optimal network size due to differences in inter‐ and intra‐migrant group externalities and subsequently coordination failures. In short, it stresses the active role incumbents take in the provision of network effects and provides theoretical evidence that self‐perpetuating and sustained immigration is not at odds with rational acting individuals. It also shows that optimally chosen migration quotas may appeal to incumbent migration cohorts and provides an alternative explanation for inter‐ and intra‐migrant group tensions.

A genetic algorithm-based methodology for optimizing multiservice convergence in a metro WDM network

We consider the multi-objective optimization of a multi-service arrayed-waveguide grating-based single-hop metro WDM network with the two conflicting objectives of maximizing throughput while minimizing delay. We develop and evaluate a genetic algorithm based methodology for finding the optimal throughput-delay tradeoff curve, the so-called Pareto-optimal frontier. Our methodology provides the network architecture (hardware) and the medium access control (MAC) protocol parameters that achieve the Pareto-optima in a computationally efficient manner. The numerical results obtained with our methodology provide the Pareto-optimal network planning and operation solutions for a wide range of traffic scenarios. The presented methodology is applicable to other networks with a similar throughput-delay tradeoff.

Coordination Failures in Network Migration

Previous migration facilitates future population moves, a phenomenon called network migration. However, thus far, network migration has been closely linked to network externalities. In contrast, this paper argues that the incumbent migration population can actively impact on future migration flows, yet fails to achieve a Pareto optimal network size due to co-ordination failures. In short, it stresses the active role incumbents take in the provision of network effects and, second, it provides theoretical evidence that self-perpetuating and sustained immigration is not at odds with rational acting individuals and must not simply be ascribed to pure network externalities.