PL
 |
EN
Preferencje help
Polish version English version Język
Widoczny [Schowaj] Abstrakt
Liczba wyników

Znaleziono wyników: 9

Liczba wyników na stronie
first rewind previous Strona / 1 next fast forward last
Wyniki wyszukiwania
help Sortuj według:

help Ogranicz wyniki do:
first rewind previous Strona / 1 next fast forward last
1
Content available remote Application of agent-based simulated annealing and tabu search procedures to solving the data reduction problem
Czarnowski I., Jędrzejowicz P.
International Journal of Applied Mathematics and Computer Science
|
2011
|
Vol. 21, no 1
57-68
EN
The problem considered concerns data reduction for machine learning. Data reduction aims at deciding which features and instances from the training set should be retained for further use during the learning process. Data reduction results in increased capabilities and generalization properties of the learning model and a shorter time of the learning process. It can also help in scaling up to large data sources. The paper proposes an agent-based data reduction approach with the learning process executed by a team of agents (A-Team). Several A-Team architectures with agents executing the simulated annealing and tabu search procedures are proposed and investigated. The paper includes a detailed description of the proposed approach and discusses the results of a validating experiment.
2
An experimental investigation of the synergetic effect of multiple agents working together in the A-Team
Barbucha D., Jędrzejowicz P.
Systems Science
|
2008
|
Vol. 34, no 2
55-62
EN
The paper aims at evaluating experimentally the existence and strength of the synergetic effect produced by a joint effort of a number of optimization agents solving instances of one of the classic combinatorial optimization problems - the vehicle routing problem. The computational experiment carried out involved JABAT environment, which is a middleware supporting the construction of the dedicated A-Team architectures. Several types of optimization agents have been constructed and tested. The main factors in the reported experiment are structure and composition of the set of agents working in parallel on solving instances of the problem at hand.
3
Content available remote A Cross - Entropy Based Population Learning Algorithm for Discrete-Continuous Scheduling with Continuous Resource Discretisation
Jędrzejowicz P., Skakovski A.
Prace Wydziału Nawigacyjnego Akademii Morskiej w Gdyni
|
2008
|
Z. 22
45-52
EN
A problem of scheduling nonpreemtable tasks on parallel identical machines under constraint on discrete resource and requiring, additionally, renewable continuous resource to minimize the schedule length is considered in the paper. A continuous resource is divisible continuously and is allocated to tasks from given intervals in amounts unknown in advance. Task processing rate depends on the allocated amount of the continuous resource. The considered problem can be solved in two steps. The first step involves generating all possible task schedules and second - finding an optimal schedule among all schedules with optimal continuous resource allocation. To eliminate time consuming optimal continuous resource allocation, a problem Teta Z with continuous resource discretisation is introduced. Because Teta Z is NP-hard a population-learning algorithm (PLA2) is proposed to tackle the problem. PLA2 belongs to the class of the population-based methods. Experiment results proved that PLA2 excels known algorithms for solving the considered problem.
PL
W pracy został rozpatrzony problem dyskretno-ciągłego szeregowania niepodzielnych zadań na równoległych identycznych maszynach, mającego na celu minimalizację długości uszeregowania z ograniczeniami nałożonymi na zasób dyskretny i dodatkowy odnawialny zasób ciągły. Zasób ciągły, podzielny w sposób ciągły, jest przydzielany do zadań z określonych przedziałów w ilościach z góry nieznanych. Szybkość wykonania zadań zależy od przydzielonej ilości zasobu ciągłego. Rozpatrywany problem można rozwiązać dwuetapowo. W pierwszym etapie należy wygenerować wszystkie możliwe uszeregowania zadań na procesorach. Drugi etap polega na znalezieniu optymalnego uszeregowania wśród wszystkich uszeregowań z optymalnym przydziałem zasobu ciągłego. W celu wyeliminowania czasochłonnej procedury optymalnego przydziału zasobu ciągłego, rozpatrzony został problem Teta Z z dyskretyzacją zasobu ciągłego. Ponieważ Teta Z jest problemem obliczeniowo NP-trudnym, do rozwiązywania został zaproponowany algorytm uczenia populacji PLA2 należący do klasy algorytmów opartych na ewolucji populacji. Eksperymenty obliczeniowe udowodniły, że PLA2 znajduje lepsze rozwiązania niż inne znane algorytmy przeznaczone do rozwiązywania dyskretno-ciągłych problemów szeregowania.
4
Content available remote An agent-based, self-tuning, population learning algorithm for the resource constrained project scheduling
Jędrzejowicz P., Ratajczak-Ropel E.
Foundations of Computing and Decision Sciences
|
2007
|
Vol. 32, No. 3
213-225
EN
The authors propose an agent-based population-learning algorithm (PLA) designed for solving the RCPSP and the MRCPSP. The paper contains problem formulation and a description of the proposed implementation of the PLA. The resulting multiple-agent system has been implemented using the JABAT environment designed with a view to facilitate development of a-teams. To validate the approach a computational experiment has been earned out. It has involved instances obtained from the available benchmark data sets. Results of the experiment show that the proposed implementation can serve as an effective tool for solving the resource-constrained project scheduling problems.
5
Content available remote PLA-based permutation-scheduling
Jędrzejowicz J., Jędrzejowicz P.
Foundations of Computing and Decision Sciences
|
2003
|
Vol. 28, No. 3
159-177
EN
The paper proposes the implementation of population learning algorithm (PLA) for solving three well-known NP-hard permutation-scheduling problems. PLA is a recently developed method belonging to the class of the population-based algorithms. One of possible applications of the PLA is solving difficult optimization problems. The first of the discussed problems involves scheduling tasks on a single machine against common due date with earliness and tardiness penalties. The second is known as the permutation flow shop problem. The third one involves scheduling tasks on a single machine with total weighted tardiness as a criterion. To evaluate the proposed implementations computational experiments have been carried. They involved solving available sets of benchmark problems and comparing the results with the optimum or best-known solutions. PLA has found better upper bounds on several benchmark instances. Experiments have also helped to identify some behavioral characteristics of the proposed algorithms.
6
Application of the population learning algorithm to training feed-forward ANN
Czarnowski I., Jędrzejowicz P.
Prace Naukowe Politechniki Warszawskiej. Elektronika
|
2002
|
z. 139
147--161
EN
The chapter investigates the application of the new metaheuristic called the population learning algorithm (PLA) to training feed-forward artificial neural networks. This chapter introduces the population learning algorithm and proposes several implementations developed with a view to training several benchmark neural networks. The approach is compared with two alternative methods of training: quick propagation and genetic programming. Computer experiments show the high effectiveness and good quality of the suggested approach.
7
Content available remote An application of population-based metaheuristics to minimizing schedule makespan
Forkiewicz M., Jędrzejowicz P.
Prace Instytutu Podstaw Informatyki Polskiej Akademii Nauk
|
2001
|
Nr 928
21-24
EN
The paper investigates application of the three population-based metaheuristics - evolution algorithm, population learning algorithm and ant colony optimization to minimizing schedule makespan. All three approaches are briefly revieved. Their implementation to schedule makespan minimization in case of independent, non-preemptable tasks and multiple processors is described. The approach is evaluated by means of computational experiment, results of which are presented and discussed.
8
Ocena efektywności algorytmu szeregowania zadań wieloprocesorowych opartego na sztucznej sieci neuronowej
Czarnowski I., Jędrzejowicz P.
Zeszyty Naukowe. Automatyka / Politechnika Śląska
|
2000
|
z. 129
51-60
PL
W pracy przedstawiono wyniki analizy efektywności algorytmu szeregowania zadań wieloprocesorowych opartego na sztucznej sieci neuronowej. Zadania, o których mowa, mogą, przykładowo, reprezentować struktury lub moduły programów tolerujących błędy. Problem szeregowania zadań wieloprocesorowych dotyczy systemów czasu rzeczywistego, od których wymaga się dużej niezawodności i bezpieczeństwa. W pracy przedstawiono przybliżony algorytm szeregowania oparty na strukturze sztucznej sieci neuronowej. Algorytm został poddany ocenie metodą eksperymentów obliczeniowych, których rezultaty również przedstawiono w niniejszej pracy.
EN
The paper proposes an artificial neural network based algorithm for scheduling multiprocessor tasks. The considered scheduling problem class is characterized by a set of multiple, identical processors and a set of multiple-processor tasks. Multiple-processor tasks have to be processed on more than one processor at a time. Decision variables include assignment of processors to tasks and size of each task. To solve the above problem an artificial neural network based algorithm is proposed. Computational complexity of the approach is analyzed and evaluated. To assess effectiveness of the algorithm computational experiment has been carried. The results obtained by using the proposed algorithm have been compared with those generated by the specially designed evolutionary algorithm and the hybrid evolutionary. Experiment results prove high effectiveness of the neural network approach.
9
Content available remote Social learning algorithm as a tool for solving some difficult scheduling problems
Jędrzejowicz P.
Foundations of Computing and Decision Sciences
|
1999
|
Vol. 24, No. 2
51-66
EN
The paper investigates application of the new adaptive memory programming techniques called SLA (social learning algorithm) to scheduling multiple variant tasks. SLA has been inspired by analogies to a social phenomenon of organized learning. Main features of both -education systems and SLA include learning stages, massive number of individuals at the beginning stages, selection of best individuals.- before entering further stages of learning, and increasing sophistication and specialization of learning as an individual progresses through stages. The paper introduces a concept of the social learning algorithm (SLA ), and presents two applications of the proposed concept to solving computationally difficult scheduling problems. Computational experiments show that SLA can produce competitive results comparing to other heuristics and metaheuristics.
first rewind previous Strona / 1 next fast forward last
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.