Time filter

Source Type

Cooperation as a problem-solving strategy is widely used to build methods addressing complex hard optimization problems. It involves a set of highly autonomous programs (agents), each implementing a particular solution method, and a cooperation scheme combining these autonomous programs into a single problem-solving strategy. Possible form of such cooperation may be based, for example, on adaptive memory methods, where partial elements of good solutions are stored and next combined to create new complete solutions. Alternative approach is based on central memory, where complete elite solutions are exchanged among various agents and/or heuristics. Moreover, cooperatively solving a task is often combined with learning mechanism, where agents adapt their behavior to the new states of environment during the process of solving the problem.The main goal of the paper is to evaluate to what extent a mode of cooperation (synchronous or asynchronous) between a number of optimization agents cooperating through sharing a central memory influences the quality of solutions while solving instances of the Vehicle Routing Problem. The investigated search modes are evaluated using a dedicated cooperative multi-agent system allowing for various modes of cooperation and with the reinforcement learning mechanism implemented in it. © 2012 Elsevier B.V. Source

Czarnowski I.,Maritime Academy of Gdynia
Pattern Recognition

Distributed learning from data is one of the typical tasks solved by distributed data-mining techniques and is seen as a fundamental computational problem. One of the approaches suitable for distributed learning is to select, by data reduction, relevant local patterns, called also prototypes, from geographically distributed databases. Next, locally selected prototypes can be moved to other sites and merged into the global knowledge model. The paper presents three agent-based population learning algorithms for distributed learning. The proposed algorithms are based on agent collaborations in distributed prototype selection processes and on agent collaborations when the learning global model is created. The basic property of the presented algorithms is that the prototypes are selected by agent-based population learning algorithm from data clusters induced at distributed sites. The main goal of the paper is to empirically compare how the way of inducing such clusters can influence the distributed learning performance. The paper investigates the agent-based population learning algorithms used to solve distributed data reduction and gives a brief discussion of the procedures for clusters initialization. Finally, computational experiment results are shown. © 2010 Elsevier Ltd. All rights reserved. Source

Presented paper shows the results of the laboratory study on the relation between the chosen malfunctions of a fuel pump and the exhaust gas composition of the marine engine. The object of research is a laboratory four-stroke diesel engine, operated at a constant speed. During the research over 50 parameters were measured with technical condition of the engine recognized as "working properly" and with simulated fuel pump malfunctions. Considered malfunctions are: fuel injection timing delay and two sets of fuel leakages in the fuel pump of one engine cylinder. The results of laboratory research confirm that fuel injection timing delay and fuel leakage in the fuel pump cause relatively small changes in thermodynamic parameters of the engine. Changes of absolute values are so small they may be omitted by marine engines operators. The measuring of the exhaust gas composition shows markedly affection with simulated malfunctions of the fuel pump. Engine operation with delayed fuel injection timing in one cylinder indicates CO2 emission increase and NOx emission decreases. CO emission increases only at high the engine loads. Fuel leakage in the fuel pump causes changes in CO emission, the increase of CO2 emission and the decrease of NOx emission. © 2014 Elsevier Ltd. All rights reserved. Source

Soszynska J.,Maritime Academy of Gdynia
International Journal of Pressure Vessels and Piping

The semi-Markov model of the system operation processes is proposed and its selected characteristics are determined. A system composed on multi-state components is considered and its reliability and risk characteristics are found. Next, the joint model of the system operation process and the system multi-state reliability is applied to the reliability and risk evaluation of the port oil pipeline transportation system. The pipeline system is described and its operation process unknown parameters are identified on the basis of real statistical data. The mean values of the pipeline system operation process unconditional sojourn times in particular operation states are found and applied to determining this process transient probabilities in these states. The piping different reliability structures in various its operation states are fixed and their conditional reliability functions on the basis of data coming from experts are approximately determined. Finally, after applying earlier estimated transient probabilities and system conditional reliability functions in particular operation states the unconditional reliability function, the mean values and standard deviations of the pipeline lifetimes in particular reliability states, risk function and the moment when the risk exceeds a critical value are found. © 2010 Elsevier Ltd. All rights reserved. Source

Gnacinski P.,Maritime Academy of Gdynia
Energy Conversion and Management

This work deals with the effect of a lowered voltage quality on the thermal loss of life and load-carrying capacity of marine induction cage machines. Results of experimental investigations and computer calculations are presented for two low power induction motors with different properties. One of them has a comparatively strongly-saturated magnetic circuit and is especially exposed to the risk of overheating under overvoltage. The other machine has a comparatively weakly-saturated magnetic circuit, and is especially sensitive to undervoltage. The induction motor lifetime expectancy is also estimated on the basis of the temperature coefficient of power quality, whose value is proportional to the windings temperature rise in induction motors especially sensitive to various power quality disturbances. The dependence of the temperature coefficient of power quality and permissible loads for induction motors supplied with voltages of lowered quality is proposed. © 2013 Elsevier Ltd. All rights reserved. Source

Discover hidden collaborations