Agency and structure: a social simulation of knowledge-intensive industries

Modern knowledge-intensive economies are complex social systems where intertwining factors are responsible for the shaping of emerging industries: the self-organising interaction patterns and strategies of the individual actors (an agency-oriented pattern) and the institutional frameworks of different innovation systems (a structure-oriented pattern). In this paper, we examine the relative primacy of the two patterns in the development of innovation networks, and find that both are important. In order to investigate the relative significance of strategic decision making by innovation network actors and the roles played by national institutional settings, we use an agent-based model of knowledge-intensive innovation networks, SKIN. We experiment with the simulation of different actor strategies and different access conditions to capital in order to study the resulting effects on innovation performance and size of the industry. Our analysis suggests that actors are able to compensate for structural limitations through strategic collaborations. The implications for public policy are outlined.     

Routing and Admission Control in General Topology Networks with Poisson Arrivals

Emerging high-speed networks will carry traffic for services such as video-on-demand and video teleconferencing that require resource reservation along the path on which the traffic is sent. High bandwidth-delay product of these networks prevents circuit rerouting, i.e., once a circuit is routed on a certain path, the bandwidth taken by this circuit remains unavailable for the duration (holding time) of this circuit. As a result, such networks will need effectiveroutingandadmission controlstrategies. Recently developed on-line routing and admission control strategies have logarithmic competitive ratios with respect to theadmission ratio(the fraction of admitted circuits). Such guarantees on performance are rather weak in the most interesting case where the rejection ratio of the optimum algorithm is very small or even 0. Unfortunately, these guarantees cannot be improved in the context of the considered models, making it impossible to use these models to identify algorithms that are going to perform well in practice. In this paper we develop routing and admission control strategies for a probabilistic model, where the requests for virtual circuits between any two points arrive according to a Poisson process and where the circuit holding times are exponentially distributed. Our model is close to the one that was developed to analyze and tune the (currently used) strategies for managing traffic in long-distance telephone networks. We strengthen this model by assuming that the rates of the Poisson processes (the “traffic matrix”) are unknown to the algorithm and are chosen by the adversary. Our strategy is competitive with respect to the expectedrejection ratio. More precisely, it achieves an expected rejection ratio of at mostR* + ?, whereR* is the optimum expected rejection ratio. The expectations are taken over the distribution of the request sequences, and, whereris the maximum fraction of an edge bandwidth that can be requested by a single circuit. Our result should be viewed in the context of the previous competitive routing and admission control strategies that requirer ≤ 1/log n, but are not able to formally analyze the (intuitively clear) relation betweenrand the performance achievable in realistic situations.     

Cascade-robustness optimization of coupling preference in interconnected networks

Recently, the robustness of interconnected networks has attracted extensive attentions, one of which is to investigate the influence of coupling preference. In this paper, the memetic algorithm (MA) is employed to optimize the coupling links of interconnected networks. Afterwards, a comparison is made between MA optimized coupling strategy and traditional assortative, disassortative and random coupling preferences. It is found that the MA optimized coupling strategy with a moderate assortative value shows an outstanding performance against cascading failures on both synthetic scale-free interconnected networks and real-world networks. We then provide an explanation for this phenomenon from a micro-scope point of view and propose a coupling coefficient index to quantify the coupling preference. Our work is helpful for the design of robust interconnected networks.     

Analysis of complex network performance and heuristic node removal strategies

Removing important nodes from complex networks is a great challenge in fighting against criminal organizations and preventing disease outbreaks. Six network performance metrics, including four new metrics, are applied to quantify networks’ diffusion speed, diffusion scale, homogeneity, and diameter. In order to efficiently identify nodes whose removal maximally destroys a network, i.e., minimizes network performance, ten structured heuristic node removal strategies are designed using different node centrality metrics including degree, betweenness, reciprocal closeness, complement-derived closeness, and eigenvector centrality. These strategies are applied to remove nodes from the September 11, 2001 hijackers’ network, and their performance are compared to that of a random strategy, which removes randomly selected nodes, and the locally optimal solution (LOS), which removes nodes to minimize network performance at each step. The computational complexity of the 11 strategies and LOS is also analyzed. Results show that the node removal strategies using degree and betweenness centralities are more efficient than other strategies.     

Multilevel-in-width training for deep neural network regression

A common challenge in regression is that for many problems, the degrees of freedom required for a high-quality solution also allows for overfitting. Regularization is a class of strategies that seek to restrict the range of possible solutions so as to discourage overfitting while still enabling good solutions, and different regularization strategies impose different types of restrictions. In this paper, we present a multilevel regularization strategy that constructs and trains a hierarchy of neural networks, each of which has layers that are wider versions of the previous network's layers. We draw intuition and techniques from the field of Algebraic Multigrid (AMG), traditionally used for solving linear and nonlinear systems of equations, and specifically adapt the Full Approximation Scheme (FAS) for nonlinear systems of equations to the problem of deep learning. Training through V-cycles then encourage the neural networks to build a hierarchical understanding of the problem. We refer to this approach as multilevel-in-width to distinguish from prior multilevel works which hierarchically alter the depth of neural networks. The resulting approach is a highly flexible framework that can be applied to a variety of layer types, which we demonstrate with both fully connected and convolutional layers. We experimentally show with PDE regression problems that our multilevel training approach is an effective regularizer, improving the generalize performance of the neural networks studied.     

Asymptotic analysis for closed multiclass queueing networks in critical usage

We consider a class of closed multiclass queueing networks containing First-Come-First-Serve (FCFS) and Infinite Server (IS) stations. These networks have a productform solution for their equilibrium probabilities. We study these networks in an asymptotic regime for which the number of customers and the service rates at the FCFS stations go to infinity with the same order. We assume that the regime is in critical usage, whereby the utilizations of the FCFS servers slowly approach one. The asymptotic distribution of the normalized queue lengths is shown to be in many cases a truncated multivariate normal distribution. Traffic conditions for which the normalized queue lengths arealmost asymptotically independent are determined. Asymptotic expansions of utilizations and expected queue lengths are presented. We show through an example how to obtain asymptotic expansions of performance measures when the networks are in mixed usage and how to apply the results to networks with finite data.Supported partially by NSF grant NCR93-04601.     

A rolling window optimization method for large-scale WCDMA base stations planning problem

The large-scale base station planning problem for wideband code division multiple access (WCDMA) wireless networks is studied in this paper. A new rolling window optimization method is presented, where the global optimization problem is decomposed into small optimization sub-problems, which are defined on a series of successive rolling windows. Effective rolling strategies are designed in the rolling optimization method based on the prediction of the interference among the base stations in the WCDMA wireless network. We show that the proposed method has the property that the global objective is non-increasing in the successive optimization procedure. Simulations are carried out to analyze the performance of the proposed optimization method, which show the importance of the rolling strategy.     

A robustness approach to international sourcing

An important advantage of the use of international sourcing networks (i.e. selection of suppliers in various countries to support the demands of the firm's international factory network) is the resulting hedging power against real exchange rate changes in the international environment. Due to the uncertainty of future real exchange rate changes, the international manager wants to develop a sourcing network that is relatively insensitive (i.e. robust) to the potential changes of the macroeconomic parameters over a planning horizon. In our paper, we formally develop arobust approach to international sourcing. This approach develops the international supplier network in a way that adequately hedges the firm's performance against the worst contingency in terms of realizable real exchange rate shocks over a planning horizon. We present an algorithm to obtain theN best robust solutions (i.e. sourcing networks) to the international sourcing problem. Some computational results on the effectiveness of the approach are provided. We also demonstrate how the approach can be used to evaluate various sourcing strategies.     

Multicriterion decision making in performance evaluation of an irrigation system

The selection of the best alternative in irrigation development strategies is examined in the multiobjective context. Sri Ram Sagar Project, a major irrigation project in India is taken for the case study. The study deals with the performance evaluation of five canal distributories by considering eight different criteria. Two Multicriterion Decision Making (MCDM) methods, namely, Multi Attribute Utility Theory (MAUT) and Stochastic extension of PROMETHEE-2 (STOPROM-2) are employed to select the best one among the five. Taguchi experimental design technique is employed to minimise the computational burden arising in the sensitivity analysis studies. The proposed methodology can serve as a model to choose the best one for formulating guide lines for improving the efficiency and performance of similar other distributories.     

Rapid improvement of stochastic networks using two-moment approximations

This paper investigates a method for rapidly improving the performance of stochastic networks. The method uses queueing network techniques based on parametric decomposition and two-moment approximations. It employs some recently devised developments to the approximations, especially in the context of closed queueing networks with synchronization stations. We describe these developments and explain how the method uses them. It appears to offer a substantial advantage in speed, and reduces the dependence on repetitive simulation runs for identifying opportunities to improve network performance. We present an example from airfield operations, in which the method evaluates the tradeoffs with respect to several strategies for enhancing airfield performance, such as the introduction of concurrent operations, the reduction of service variability, and the increase in airfield capacity.     

Reduced RLT representations for nonconvex polynomial programming problems

This paper explores equivalent, reduced size Reformulation-Linearization Technique (RLT)-based formulations for polynomial programming problems. Utilizing a basis partitioning scheme for an embedded linear equality subsystem, we show that a strict subset of RLT defining equalities imply the remaining ones. Applying this result, we derive significantly reduced RLT representations and develop certain coherent associated branching rules that assure convergence to a global optimum, along with static as well as dynamic basis selection strategies to implement the proposed procedure. In addition, we enhance the RLT relaxations with v-semidefinite cuts, which are empirically shown to further improve the relative performance of the reduced RLT method over the usual RLT approach. We present computational results for randomly generated instances to test the different proposed reduction strategies and to demonstrate the improvement in overall computational effort when such reduced RLT mechanisms are employed.     

A Reduced Load Approximation Accounting for Link Interactions in a Loss Network

This paper is concerned with evaluating the performance of loss networks. Accurate determination of loss network performance can assist in the design and dimensioning of telecommunications networks. However, exact determination can be diffcult and generally cannot be done in reasonable time. For these reasons there is much interest in developing fast and accurate approximations. We develop a reduced load approximation that improves on the famous Erlang fixed point approximation (EFPA) in a variety of circumstances. We illustrate our results with reference to a range of networks for which the EFPA may be expected to perform badly.     

Artificial neural networks in bankruptcy prediction: General framework and cross-validation analysis

In this paper, we present a general framework for understanding the role of artificial neural networks (ANNs) in bankruptcy prediction. We give a comprehensive review of neural network applications in this area and illustrate the link between neural networks and traditional Bayesian classification theory. The method of cross-validation is used to examine the between-sample variation of neural networks for bankruptcy prediction. Based on a matched sample of 220 firms, our findings indicate that neural networks are significantly better than logistic regression models in prediction as well as classification rate estimation. In addition, neural networks are robust to sampling variations in overall classification performance.     

On optimizing CSMA for wide area ad hoc networks

The recent deployment of data-rich smart phones has led to a fresh impetus for understanding the performance of wide area ad hoc networks. The most popular medium access mechanism for such ad hoc networks is CSMA/CA with RTS/CTS. In CSMA-like mechanisms, spatial reuse is achieved by implementing energy-based guard zones. We consider the problem of simultaneously scheduling the maximum number of links that can achieve a given signal to interference ratio (SIR). In this paper, using tools from stochastic geometry, we study and maximize the medium access probability of a typical link. Our contributions are two-fold: (i) We show that a simple modification to the RTS/CTS mechanism, viz., changing the receiver yield decision from an energy-level guard zone to an SIR guard zone, leads to performance gains; and (ii) We show that this combined with a simple modification to the transmit power level—setting it inversely proportional to the square root of the link gain—leads to significant improvements in network throughput. Further, this simple power-level choice is no worse than a factor of two away from optimal over the class of all “local” power level selection strategies for fading channels, and further is optimal in the non-fading case. The analysis relies on an extension of the Matérn hard core point process which allows us to quantify both these SIR guard zones and this power control mechanism.     

On the asymptotic optimality of the cμ-rule in queueing networks

Although the cμ-rule is proven to be exactly or asymptotically optimal in various parallel queueing networks, its performance in non-parallel queueing networks is relatively unexplored. We study the performance of the cμ-rule (that is, the performance of implementing the cμ-rule at all servers) in non-parallel queueing networks. We numerically show that the cμ-rule can perform poorly even in a simple tandem queue with two job types. We derive conditions under which the cμ-rule is asymptotically optimal in diffusion scale.     

Pairwise-stability and Nash equilibria in network formation

Suppose that individual payoffs depend on the network connecting them. Consider the following simultaneous move game of network formation: players announce independently the links they wish to form, and links are formed only under mutual consent. We provide necessary and sufficient conditions on the network link marginal payoffs such that the set of pairwise stable, pairwise-Nash and proper equilibrium networks coincide, where pairwise stable networks are robust to one-link deviations, while pairwise-Nash networks are robust to one-link creation but multi-link severance. Under these conditions, proper equilibria in pure strategies are fully characterized by one-link deviation checks. We thank William Thomson, an associate editor and two anonymous referees for their suggestions that led to substantial improvements. We also thank Sjaak Hurkens, Bettina Klaus, Jordi Massó and Giovanni Neglia for helpful conversations. The first author gratefully acknowledges the financial support from the Spanish Ministry of Education and FEDER through grant SEJ2005-01481/ECON, the Fundación BBVA and the Barcelona Economics Program of XREA. The second author is grateful to the Netherlands Organization for Scientific Research (NWO) for its support under grant VIDI-452-06-013.     

Production systems with interruptions, arbitrary topology and finite buffers

We consider a production system with finite buffers and arbitrary topology where service time is subject to interruptions in one of three ways, viz. machine breakdown, machine vacations or Npolicy. We develop a unified approximation (analytical) methodology to calculate the throughput of the system using queueing networks together with decomposition, isolation and expansion techniques. The methodology is rigorously tested covering a large experimental region. Orthogonal arrays are used to design the experiments in order to keep the number of experiments manageable. The results obtained using the approximation methodology are compared to the simulation results. The ttests carried out to investigate the differences between the two results show that they are statistically insignificant. Finally, we test the methodology by applying it to several arbitrary topology networks. The results show that the performance of the approximation methodology is consistent, robust and produces excellent results in a variety of experimental conditions.     

Optimization of Optimal Power Flow Problems

Optimal power flow problems arise in the context of the optimization and secure exploitation of electrical power in alternating current (AC) networks. This optimization problem evaluates the flow on each line and to ensure that they are under line thermal limits. To improve the reliability of the power supply, a secure network is necessary, i.e., it has to be able to cope with some contingencies. Nowadays high performance solution methods, that are based on nonlinear programming algorithms, search for an optimal state while considering certain contingencies. According to the number of contingencies the problem size increases linearly. As the base case can already be large-scaled, the optimization time computation increases quickly. Parallelization seems to be a good way to solve quickly this kind of problem. This paper considers the minimization of an objective function and at least two constraints at each node. This optimization problem is solved by IPOPT, an interior point method, coupled with ADOL-C, an algorithmic differentiation tool, and MA27, a linear solver. Several results on employed parallelizing strategies will be discussed. (© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim)     

Optimal complex networks spontaneously emerge when information transfer is maximized at least expense: A design perspective

Complex networks with multiple nodes and diverse interactions among them are ubiquitous. We suggest that optimal networks spontaneously emerge when “information” transfer is maximized at the least expense. We support our hypothesis by evolving optimal topologies for a particle swarm optimizer (PSO), a population‐based stochastic algorithm. Results suggest that (1) an optimum topology emerges at the phase transition when connectivity is high enough to transfer information but low enough to prevent premature convergence, and (2) Small World (SW) networks are a compromise between higher performance and resistance to mutation. The graph characteristics of the optimal PSO networks in the SW regime are similar to that of the visual cortices of cat and macaque, thereby suggesting similar design principles. © 2006 Wiley Periodicals, Inc. Complexity 11:26–35, 2006     

Service rate control of closed Jackson networks from game theoretic perspective

Game theoretic analysis of queueing systems is an important research direction of queueing theory. In this paper, we study the service rate control problem of closed Jackson networks from a game theoretic perspective. The payoff function consists of a holding cost and an operating cost. Each server optimizes its service rate control strategy to maximize its own average payoff. We formulate this problem as a non-cooperative stochastic game with multiple players. By utilizing the problem structure of closed Jackson networks, we derive a difference equation which quantifies the performance difference under any two different strategies. We prove that no matter what strategies the other servers adopt, the best response of a server is to choose its service rates on the boundary. Thus, we can limit the search of equilibrium strategy profiles from a multidimensional continuous polyhedron to the set of its vertex. We further develop an iterative algorithm to find the Nash equilibrium. Moreover, we derive the social optimum of this problem, which is compared with the equilibrium using the price of anarchy. The bounds of the price of anarchy of this problem are also obtained. Finally, simulation experiments are conducted to demonstrate the main idea of this paper.