Biblio
Swarm Intelligence (SI) algorithms, such as Fish School Search (FSS), are well known as useful tools that can be used to achieve a good solution in a reasonable amount of time for complex optimization problems. And when problems increase in size and complexity, some increase in population size or number of iterations might be needed in order to achieve a good solution. In extreme cases, the execution time can be huge and other approaches, such as parallel implementations, might help to reduce it. This paper investigates the relation and trade off involving these three aspects in SI algorithms, namely population size, number of iterations, and problem complexity. The results with a parallel implementations of FSS show that increasing the population size is beneficial for finding good solutions. However, we observed an asymptotic behavior of the results, i.e. increasing the population over a certain threshold only leads to slight improvements.
Interval uncertainty can cause uncontrollable variations in the objective and constraint values, which could seriously deteriorate the performance or even change the feasibility of the optimal solutions. Robust optimization is to obtain solutions that are optimal and minimally sensitive to uncertainty. In this paper, a sequential multi-objective robust optimization (MORO) approach based on support vector machines (SVM) is proposed. Firstly, a sequential optimization structure is adopted to ease the computational burden. Secondly, SVM is used to construct a classification model to classify design alternatives into feasible or infeasible. The proposed approach is tested on a numerical example and an engineering case. Results illustrate that the proposed approach can reasonably approximate solutions obtained from the existing sequential MORO approach (SMORO), while the computational costs are significantly reduced compared with those of SMORO.
In video surveillance, face recognition (FR) systems seek to detect individuals of interest appearing over a distributed network of cameras. Still-to-video FR systems match faces captured in videos under challenging conditions against facial models, often designed using one reference still per individual. Although CNNs can achieve among the highest levels of accuracy in many real-world FR applications, state-of-the-art CNNs that are suitable for still-to-video FR, like trunk-branch ensemble (TBE) CNNs, represent complex solutions for real-time applications. In this paper, an efficient CNN architecture is proposed for accurate still-to-video FR from a single reference still. The CCM-CNN is based on new cross-correlation matching (CCM) and triplet-loss optimization methods that provide discriminant face representations. The matching pipeline exploits a matrix Hadamard product followed by a fully connected layer inspired by adaptive weighted cross-correlation. A triplet-based training approach is proposed to optimize the CCM-CNN parameters such that the inter-class variations are increased, while enhancing robustness to intra-class variations. To further improve robustness, the network is fine-tuned using synthetically-generated faces based on still and videos of non-target individuals. Experiments on videos from the COX Face and Chokepoint datasets indicate that the CCM-CNN can achieve a high level of accuracy that is comparable to TBE-CNN and HaarNet, but with a significantly lower time and memory complexity. It may therefore represent the better trade-off between accuracy and complexity for real-time video surveillance applications.
Today the technology advancement in communication technology permits a malware author to introduce code obfuscation technique, for example, Application Programming Interface (API) hook, to make detecting the footprints of their code more difficult. A signature-based model such as Antivirus software is not effective against such attacks. In this paper, an API graph-based model is proposed with the objective of detecting hook attacks during malicious code execution. The proposed model incorporates techniques such as graph-generation, graph partition and graph comparison to distinguish a legitimate system call from malicious system call. The simulation results confirm that the proposed model outperforms than existing approaches.
Location-Based Service (LBS) becomes increasingly important for our daily life. However, the localization information in the air is vulnerable to various attacks, which result in serious privacy concerns. To overcome this problem, we formulate a multi-objective optimization problem with considering both the query probability and the practical dummy location region. A low complexity dummy location selection scheme is proposed. We first find several candidate dummy locations with similar query probabilities. Among these selected candidates, a cloaking area based algorithm is then offered to find K - 1 dummy locations to achieve K-anonymity. The intersected area between two dummy locations is also derived to assist to determine the total cloaking area. Security analysis verifies the effectiveness of our scheme against the passive and active adversaries. Compared with other methods, simulation results show that the proposed dummy location scheme can improve the privacy level and enlarge the cloaking area simultaneously.
Hierarchical approaches for representation learning have the ability to encode relevant features at multiple scales or levels of abstraction. However, most hierarchical approaches exploit only the last level in the hierarchy, or provide a multiscale representation that holds a significant amount of redundancy. We argue that removing redundancy across the multiple levels of abstraction is important for an efficient representation of compositionality in object-based representations. With the perspective of feature learning as a data compression operation, we propose a new greedy inference algorithm for hierarchical sparse coding. Convolutional matching pursuit with a L0-norm constraint was used to encode the input signal into compact and non-redundant codes distributed across levels of the hierarchy. Simple and complex synthetic datasets of temporal signals were created to evaluate the encoding efficiency and compare with the theoretical lower bounds on the information rate for those signals. Empirical evidence have shown that the algorithm is able to infer near-optimal codes for simple signals. However, it failed for complex signals with strong overlapping between objects. We explain the inefficiency of convolutional matching pursuit that occurred in such case. This brings new insights about the NP-hard optimization problem related to using L0-norm constraint in inferring optimally compact and distributed object-based representations.
Complex safety-critical devices require dependable communication. Dependability includes confidentiality and integrity as much as safety. Encrypting gateways with demilitarized zones, Multiple Independent Levels of Security architectures and the infamous Air Gap are diverse integration patterns for safety-critical infrastructure. Though resource restricted embedded safety devices still lack simple, certifiable, and efficient cryptography implementations. Following the recommended formal methods approach for safety-critical devices, we have implemented proven cryptography algorithms in the qualified model based language Scade as the Safety Leveraged Implementation of Data Encryption (SLIDE) library. Optimization for the synchronous dataflow language is discussed in the paper. The implementation for public-key based encryption and authentication is evaluated for real-world performance. The feasibility is shown by execution time benchmarks on an industrial safety microcontroller platform running a train control safety application.
Tactical networks are generally simple ad-hoc networks in design, however, this simple design often gets complicated, when heterogeneous wireless technologies have to work together to enable seamless multi-hop communications across multiple sessions. In recent years, there has been some significant advances in computational, radio, localization, and networking te, and session's rate i.e., aggregate capacity averaged over a 4-time-slot frame)chnologies, which motivate a clean slate design of the control plane for multi-hop tactical wireless networks. In this paper, we develop a global network optimization framework, which characterizes the control plane for multi-hop wireless tactical networks. This framework abstracts the underlying complexity of tactical wireless networks and orchestrates the the control plane functions. Specifically, we develop a cross-layer optimization framework, which characterizes the interaction between the physical, link, and network layers. By applying the framework to a throughput maximization problem, we show how the proposed framework can be utilized to solve a broad range of wireless multi-hop tactical networking problems.
End-hopping is an effective component of Moving Target Defense (MTD) by randomly hopping network configuration of host, which is a game changing technique against cyber-attack and can interrupt cyber kill chain in the early stage. In this paper, a novel end-hopping model, Multi End-hopping (MEH), is proposed to exploit the full potentials of MTD techniques by hosts cooperating with others to share possible configurable space (PCS). And an optimization method based on cooperative game is presented to make hosts form optimal alliances against reconnaissance, scanning and blind probing DoS attack. Those model and method confuse adversaries by establishing alliances of hosts to enlarge their PCS, which thwarts various malicious scanning and mitigates probing DoS attack intensity. Through simulations, we validate the correctness of MEH model and the effectiveness of optimization method. Experiment results show that the proposed model and method increase system stable operational probability while introduces a low overhead in optimization.
Control plane distribution on Software Defined Networking enhances security, performance and scalability of the network. In this paper, we propose an efficient architecture for distribution of controllers. The main contributions of the proposed architecture are: i) A controller distributed areas to ensure security, performance and scalability of the network; ii) A single database maintained by a designated controller to provide consistency to the control plane; iii) An optimized heuristic for locating controllers to reduce latency in the control plane; iv) A resilient mechanism of choosing the designated controller to ensure the proper functioning of the network, even when there are failures. A prototype of the proposal was implemented and the placement heuristic was analyzed in real topologies. The results show that connectivity is maintained even in failure scenarios. Finally, we show that the placement optimization reduces the average latency of controllers. Our proposed heuristic achieves a fair distribution of controllers and outperforms the network resilience of other heuristics up to two times better.
This paper proposes a context-aware, graph-based approach for identifying anomalous user activities via user profile analysis, which obtains a group of users maximally similar among themselves as well as to the query during test time. The main challenges for the anomaly detection task are: (1) rare occurrences of anomalies making it difficult for exhaustive identification with reasonable false-alarm rate, and (2) continuously evolving new context-dependent anomaly types making it difficult to synthesize the activities apriori. Our proposed query-adaptive graph-based optimization approach, solvable using maximum flow algorithm, is designed to fully utilize both mutual similarities among the user models and their respective similarities with the query to shortlist the user profiles for a more reliable aggregated detection. Each user activity is represented using inputs from several multi-modal resources, which helps to localize anomalies from time-dependent data efficiently. Experiments on public datasets of insider threats and gesture recognition show impressive results.
Network systems, such as transportation systems and water supply systems, play important roles in our daily life and industrial production. However, a variety of disruptive events occur during their life time, causing a series of serious losses. Due to the inevitability of disruption, we should not only focus on improving the reliability or the resistance of the system, but also pay attention to the ability of the system to response timely and recover rapidly from disruptive events. That is to say we need to pay more attention to the resilience. In this paper, we describe two resilience models, quotient resilience and integral resilience, to measure the final recovered performance and the performance cumulative process during recovery respectively. Based on these two models, we implement the optimization of the system recovery strategies after disruption, focusing on the repair sequence of the damaged components and the allocation scheme of resource. The proposed research in this paper can serve as guidance to prioritize repair tasks and allocate resource reasonably.
We present an optimization approach that can be employed to calculate the globally optimal segmentation of a 2-D magnetic system into uniformly magnetized pieces. For each segment, the algorithm calculates the optimal shape and the optimal direction of the remanent flux density vector, with respect to a linear objective functional. We illustrate the approach with results for magnet design problems from different areas, such as a permanent magnet electric motor, a beam-focusing quadrupole magnet for particle accelerators, and a rotary device for magnetic refrigeration.
In this work we put forward our novel approach using graph partitioning and Micro-Community detection techniques. We firstly use algebraic connectivity or Fiedler Eigenvector and spectral partitioning for community detection. We then used modularity maximization and micro level clustering for detecting micro-communities with concept of community energy. We run micro-community clustering algorithm recursively with modularity maximization which helps us identify dense, deeper and hidden community structures. We experimented our MicroCommunity Clustering (MCC) algorithm for various types of complex technological and social community networks such as directed weighted, directed unweighted, undirected weighted, undirected unweighted. A novel fact about this algorithm is that it is scalable in nature.
Modern smart surveillance systems can not only record the monitored environment but also identify the targeted objects and detect anomaly activities. These advanced functions are often facilitated by deep neural networks, achieving very high accuracy and large data processing throughput. However, inappropriate design of the neural network may expose such smart systems to the risks of leaking the target being searched or even the adopted learning model itself to attackers. In this talk, we will present the security challenges in the design of smart surveillance systems. We will also discuss some possible solutions that leverage the unique properties of emerging nano-devices, including the incurred design and performance cost and optimization methods for minimizing these overheads.
A waste heat recovery system (WHRS) on a process with variable output, is an example of an intermittent renewable process. WHRS recycles waste heat into usable energy. As an example, waste heat produced from refrigeration can be used to provide hot water. However, consistent with most intermittent renewable energy systems, the likelihood of waste heat availability at times of demand is low. For this reason, the WHRS may be coupled with a hot water reservoir (HWR) acting as the energy storage system that aims to maintain desired hot water temperature Td (and therefore energy) at time of demand. The coupling of the WHRS and the HWR must be optimised to ensure higher efficiency given the intermittent mismatch of demand and heat availability. Efficiency of an WHRS can be defined as achieving multiple objectives, including to minimise the need for back-up energy to achieve Td, and to minimise waste heat not captured (when the reservoir volume Vres is too small). This paper investigates the application of a Multi Objective Evolutionary Algorithm (MOEA) to optimise the parameters of the WHRS, including the Vres and depth of discharge (DoD), that affect the WHRS efficiency. Results show that one of the optimum solutions obtained requires the combination of high Vres, high DoD, low water feed in rate, low power external back-up heater and high excess temperature for the HWR to ensure efficiency of the WHRS.
The Critical Node Detection Problem (CNDP) is a well-known NP-complete, graph-theoretical problem with many real-world applications in various fields such as social network analysis, supply-chain network analysis, transport engineering, network immunization, and military strategic planning. We present the first parallel algorithms for CNDP solving in general, and for fast, approximated CND on GPU and in the cloud in particular. Finally, we discuss results of our experimental performance analysis of these solutions.
In this paper, we study the security and system congestion in a risk-based checkpoint screening system with two kinds of inspection queues, named as Selectee Lanes and Normal Lanes. Based on the assessed threat value, the arrival crossing the security checkpoints is classified as either a selectee or a non-selectee. The Selectee Lanes with enhanced scrutiny are used to check selectees, while Normal Lanes are used to check non-selectees. The goal of the proposed modelling framework is to minimize the system congestion under the constraints of total security and limited budget. The system congestion of the checkpoint screening system is determined through a steady-state analysis of multi-server queueing models. By solving an optimization model, we can determine the optimal threshold for differentiating the arrivals, and determine the optimal number of security devices for each type of inspection queues. The analysis conducted in this study contributes managerial insights for understanding the operation and system performance of such risk-based checkpoint screening systems.
Security breaches and attacks are becoming a more critical and, simultaneously, a challenging problems for many firms in networked supply chains. A game theory-based model is developed to investigate how interdependent feature of information security risk influence the optimal strategy of firms to invest in information security. The equilibrium levels of information security investment under non-cooperative game condition are compared with socially optimal solutions. The results show that the infectious risks often induce firms to invest inefficiently whereas trust risks lead to overinvest in information security. We also find that firm's investment may not necessarily monotonous changes with infectious risks and trust risks in a centralized case. Furthermore, relative to the socially efficient level, firms facing infectious risks may invest excessively depending on whether trust risks is large enough.
In this paper, an optimization model of automobile supply chain network with risks under fuzzy price is put forward. The supply chain network is composed of component suppliers, plants, and distribution centers. The total costs of automobile supply chain consist of variable costs, fixed costs, and transportation costs. The objective of this study is to minimize the risks of total profits. In order to deal with this model, this paper puts forward an approximation method to transform a continuous fuzzy problem into discrete fuzzy problem. The model is solved using Cplex 12.6. The results show that Cplex 12.6 can perfectly solve this model, the expected value and lower semi-variance of total profits converge with the increasing number of discretization points, the structure of automobile supply chain network keeps unchanged with the increasing number of discretization points.
Next generation cellular networks will provide users better experiences by densely deploying smaller cells, which results in more complicated interferences environment. In order to coordinate interference, power control for uplink is particularly challenging due to random locations of uplink transmitter and dense deployment. In this paper, we address the uplink fractional power control (FPC) optimization problem from network optimization perspective. The relations between FPC parameters and network KPIs (Key Performance Indicators) are investigated. Rather than considering any single KPI in conventional approaches, multi-KPI optimization problem is formulated and solved. By relaxing the discrete optimization problem to a continuous one, the gradients of multiple KPIs with respect to FPC parameters are derived. The gradient enables efficiently searching for optimized FPC parameters which is particularly desirable for dense deployment of large number of cells. Simulation results show that the proposed scheme greatly outperforms the traditional one, in terms of network mean load, call drop & block ratio, and convergence speed.
We consider the problem of robust on-line optimization of a class of continuous-time nonlinear systems by using a discrete-time controller/optimizer, interconnected with the plant in a sampled-data structure. In contrast to classic approaches where the controller is updated after a fixed sufficiently long waiting time has passed, we design an event-based mechanism that triggers the control action only when the rate of change of the output of the plant is sufficiently small. By using this event-based update rule, a significant improvement in the convergence rate of the closed-loop dynamics is achieved. Since the closed-loop system combines discrete-time and continuous-time dynamics, and in order to guarantee robustness and semi-continuous dependence of solutions on parameters and initial conditions, we use the framework of hybrid set-valued dynamical systems to analyze the stability properties of the system. Numerical simulations illustrate the results.
The paper presents a joint optimization algorithm for coverage and capacity in heterogeneous cellular networks. A joint optimization objective related to capacity loss considering both coverage hole and overlap area based on power density distribution is proposed. The optimization object is a NP problem due to that the adjusting parameters are mixed with discrete and continuous, so the bacterial foraging (BF) algorithm is improved based on network performance analysis result to find a more effective direction than randomly selected. The results of simulation show that the optimization object is feasible gains a better effect than traditional method.