A performance study of multiprocessor task scheduling algorithms

Multiprocessor task scheduling is an important and computationally difficult problem. A large number of algorithms were proposed which represent various tradeoffs between the quality of the solution and the computational complexity and scalability of the algorithm. Previous comparison studies have frequently operated with simplifying assumptions, such as independent tasks, artificially generated problems or the assumption of zero communication delay. In this paper, we propose a comparison study with realistic assumptions. Our target problems are two well known problems of linear algebra: LU decomposition and Gauss–Jordan elimination. Both algorithms are naturally parallelizable but have heavy data dependencies. The communication delay will be explicitly considered in the comparisons. In our study, we consider nine scheduling algorithms which are frequently used to the best of our knowledge: min–min, chaining, A^*, genetic algorithms, simulated annealing, tabu search, HLFET, ISH, and DSH with task duplication. Based on experimental results, we present a detailed analysis of the scalability, advantages and disadvantages of each algorithm.

Active time scheduling for rechargeable sensor networks

Recent progress in energy harvesting technologies made it possible to build sensor networks with rechargeable nodes which target an indefinitely long operation. In these networks, the goal of energy management is to allocate the available energy such that the important performance metrics, such as the number of detected threats, are maximized. As the harvested energy is not sufficient for continuous operation, the scheduling of the active and inactive time is one of the main components of energy management. The active time scheduling protocols need to maintain the energy equilibrium of the nodes, while considering the uncertainties of the energy income, which is strongly influenced by the weather, and the energy expenditures, which are dependent on the behavior of the targets. In this paper, we describe and experimentally compare three active time scheduling protocols: (a) static active time, (b) dynamic active time based on a multi-parameter heuristic and (c) utility-based uniform sensing. We show that protocols which take into consideration the probabilistic models of the energy income and expenditure and can dynamically adapt to changes in the environment, can provide a significant performance advantage.     

Multi-Modal Dialog Scene Detection Using Hidden Markov Models for Content-Based Multimedia Indexing

A class of audio-visual data (fiction entertainment: movies, TV series) is segmented into scenes, which contain dialogs, using a novel hidden Markov model-based (HMM) method. Each shot is classified using both audio track (via classification of speech, silence and music) and visual content (face and location information). The result of this shot-based classification is an audio-visual token to be used by the HMM state diagram to achieve scene analysis. After simulations with circular and left-to-right HMM topologies, it is observed that both are performing very good with multi-modal inputs. Moreover, for circular topology, the comparisons between different training and observation sets show that audio and face information together gives the most consistent results among different observation sets.     

Analyzing IT maintenance outsourcing decision from a knowledge management perspective

IT outsourcing is a complex and opaque decision problem. Managers facing a decision about IT outsourcing have difficulty in framing what needs to be thought about further in their discourses. Framing is one of the most crucial steps of human decision making and needs to be assisted to better understand a decision situation. In this research, we examine a number of decision primitives in the context of an IT outsourcing decision situation. We demonstrate how the decision primitives can be employed so that managers can probe deep to better understand a decision situation and to establish a decision basis. In the organizational setting, we exemplify the use of the decision primitives in relation to the perceived outsourcing implications for the managers looking for assistance in accommodating a knowledge management perspective on IT outsourcing. Consequently, we induce insight and a guideline on how to use knowledge management for effective outsourcing in one of the leading financial institutes in Europe.

Coordinating metaheuristic agents with swarm intelligence

Coordination of multi agent systems remains as a problem since there is no prominent method suggests any universal solution. Metaheuristic agents are specific implementations of multi-agent systems, which imposes working together to solve optimisation problems using metaheuristic algorithms. An idea for coordinating metaheuristic agents borrowed from swarm intelligence is introduced in this paper. This swarm intelligence-based coordination framework has been implemented as swarms of simulated annealing agents collaborated with particle swarm optimization for multidimensional knapsack problem. A comparative performance analysis is also reported highlighting that the implementation has produced much better results than the previous works.     

The hydro-abrasive erosion wear behavior of duplex-treated surfaces of AISI H13 tool steel

The influence of duplex surface treatments consisting of a DC-pulsed plasma nitriding process and subsequent coatings of CrN and TiAlN deposited by physical vapor deposition(PVD)on AISI H13 tool steel was studied in this article.The treated samples were characterized using metallographic techniques,SEM,EDS,and microhardness methods.Hydro-abrasive erosion wear tests were performed in a specifically designed wear tester in which the samples were rotated in a wear tank containing a mixture of distilled water and ceramic abrasive chips with a fixed rotational speed.The wear rates caused by the abrasive particle impacts were assessed based on accumulated weight loss measurements.The worn surfaces were also characterized using optical microscopy,SEM,and EDS.Microhardness measurements indicated a significant increase in the surface hardness of the duplex-treated samples.The surfaces of the samples with the TiAlN coating were approximately 15 times harder than that of the untreated samples and 3 times that of the plasma nitrided samples.Hydro-abrasive erosion wear results showed that the duplex surface treatments,especially the CrN coating,displayed the highest erosion wear resistance.     

Electrophoretic deposition of nano-sized BaTiO3

In this study, a non-aqueous method in a simple one pot reaction process was employed to synthesize nano-sized BaTiO₃ particles and then electrophoretic deposition technique was employed for thin film coatings. In the first step of the preparation, metallic barium is directly dissolved in benzyl alcohol at slightly elevated temperatures. Then titanium isopropoxide was added following by a solvothermal treatment. At the end of the reaction, nearly spherical BaTiO₃ nanoparticles were obtained typically 5 nm in diameter. After establishing the stability of the BaTiO₃ suspension in ethanol, electrophoretic deposition process was performed without any additional operation. Alumina with platinum plating was used as substrate. To achieve the optimal process parameters, various voltages were applied by altering the cathode to anode distance as well as deposition time. High voltages application was possible without causing hydrolysis, because of the non-aqueous ethanol medium with higher surface charge of the nanoparticles. The deposited surface coatings were dried in air and sintered at various temperatures. SEM, EDX and XRD analysis were employed for the investigation of the coating.     

Stochastic Model for Landfill Gas Transport and Energy Recovery

Predicting the amount of landfill gas (LFG) that will be recovered at a sanitary landfill is generally associated with a high level of uncertainty, which is primarily due to the uncertainty in the definition of the parameters that control the LFG generation rate and LFG transport. To quantify these uncertainties, a three-dimensional stochastic model for the generation and transport of LFG is proposed. Using Monte Carlo simulations, multiple realizations of key input parameters are generated. For each realization, LFG transport is simulated and then used to evaluate probabilistically the rates and efficiency of energy recovery. For demonstration, the stochastic model is applied to the Kemerburgaz landfill in Istanbul, Turkey. Uncertainty in the definition of three key parameters, namely: the LFG production rate, the waste gas permeability and the soil cover gas permeability were accounted for. Modeling results suggest that the collection system is sufficient to capture most of the generated gas, but that uncertainty in the factors controlling LFG production is the main source of uncertainty in the amount of energy that will be recovered.     

Adaptive Q control for tapping-mode nanoscanning using a piezoactuated bimorph probe

A new approach, called adaptive Q control, for tapping-mode atomic force microscopy (AFM) is introduced and implemented on a homemade AFM setup utilizing a laser Doppler vibrometer and a piezoactuated bimorph probe. In standard Q control, the effective Q factor of the scanning probe is adjusted prior to the scanning depending on the application. However, there is a trade-off in setting the effective Q factor of an AFM probe. The Q factor is either increased to reduce the tapping forces or decreased to increase the maximum achievable scan speed. Realizing these two benefits simultaneously using standard Q control is not possible. In adaptive Q control, the Q factor of the probe is set to an initial value as in standard Q control, but then modified on the fly during scanning when necessary to achieve this goal. In this article, we present the basic theory behind adaptive Q control, the electronics enabling the online modification of the probe's effective Q factor, and the results of the experiments comparing three different methods: scanning (a) without Q control, (b) with standard Q control, and (c) with adaptive Q control. The results show that the performance of adaptive Q control is superior to the other two methods.     

A simulated annealing algorithm for multi-agent systems: a job-shop scheduling application

In this paper, a parallel implementation of the modular simulated annealing algorithm for classical job-shop scheduling is presented. The implementation is for a multi agent system running on the distributed resource machine, which is a novel, scalable, distributed virtual machine based on Java technology. The problems tackled are well known, difficult benchmarks, widely used to measure the efficiency of metaheuristics with respect to both the quality of the solutions and the central processing unit time. The empirical results obtained show that the method proposed is successful in comparison with a sequential version of modular simulated annealing algorithm and other methods described in the literature.