Prof. Adam Janiak
Wroclaw University of Technology, Poland

Title of talk

On Scheduling Problems with an Intelligent Use of the Learning Effect (Authors: Janiak A., Janiak W., Rudek R.)

Abstract

The talk is devoted to scheduling problems with an intelligent use of the learning effect, which is understood as a process of an acquiring experience that increases the efficiency of a processor. A measurable result of this effect is decreasing of job processing times. The existence of this phenomenon in many intelligent systems is undoubted, thus it is perceived as a worthwhile to be taken into consideration.

First, a short survey of the results concerning scheduling problems with the learning effect is provided. In particular, the existing models of the experience are presented along with a discussion on different shapes of the learning curve. We analyze scheduling problems in a single processor environment with the following minimization objectives: the makespan with job release dates, the maximum lateness and the number of late jobs. We prove that these problems become strongly NP-hard with the position dependent learning effect and stepwise learning curves. For this group of problems, we provide fast heuristic algorithms with their worst case analysis. Next, we show that an optimal solution of a two-machine flowshop problem with the makespan minimization does not have to be the `permutation' schedule if the position dependent learning effect is taken into consideration. We prove that the permutation versions of this problem with stepwise or piecewise-linear learning curves are strongly NP-hard. It is shown that permutation and non-permutation problems are NP-hard even if the learning effect, in a form of a step learning curve, characterizes only one machine.

Finally, we focus on a single machine makespan minimization problem with the experience dependent learning models. We prove that this problem is NP-hard even if the experience provided by each job is equal to its normal processing time and the learning curve is S-shaped, piecewise-linear or stepwise. To solve this problem, we prove some eliminating properties that are used to construct an efficient branch and bound algorithm. Polynomially solvable cases of the considered problems are also provided.

Short bio

Adam Janiak received the M.Eng. and Ph.D. degrees from the Wroclaw University of Technology, Wroclaw, Poland, in 1972 and 1977, respectively, and the Dr.Sc. degree from the Warsaw University of Technology, Warsaw, Poland, in 1992.

He received the Professor title in 1999 from President of Poland. He was invited as a Visiting Professor to universities in Australia, Canada, Germany, Hong Kong, Israel, New Zealand, Thailand, China, Spain, USA, Greece, Great Britain, Holland and France. Currently he is a Full Professor in computer science and operations research of industrial engineering areas with the Institute of Computer Engineering, Control and Robotics, Wroclaw University of Technology, where he is the Head of the Department of Artificial Intelligence and Algorithms Design. He has authored three books and more than 200 papers in edited books, international journals, and conference proceedings (including 57 publications in journals from ISI Master Journal List). His papers were cited over 300 times (according to data from ISI journals database). His research interests include sequencing and scheduling problems with classical and generalized models of operations in computer and manufacturing systems, resource allocation problems, complexity theory, and theory of algorithms (physical design automation of VLSI).

Prof. Janiak is a corresponding member of the Polish Academy of Sciences, a vice-president of the Computer Science Committee of the Polish Academy of Sciences and a head of the panel: “Computer Methods in Science” of the Polish Research Council. He is the expert of the State Accreditation Committee. He has served on the program committees for several international conferences on operations research and is a regular reviewer for a number of prestigious journals and conferences. He is an Associated Editor for International Journal of Applied Mathematics and Computer Science, Decision Making in Manufacturing and Services, Recent Patents on Computer Science and Book Series Computational Intelligence and its Applications.

Prof. Colin Fyfe
University of the West of Scotland

Title of talk

Data Mining and Visualisation

Abstract:

     One of the major tasks today is to create information from data. People are very good at pattern recognition; we are far more robust pattern matchers than any current computer programs. Increasingly however, we are dealing with high dimensional (and often high volume) data so to gain intuitions about a data set, we often project data onto low dimensional manifolds. One question which arises then, is what projections to low dimensional manifolds are best in order to present the projected data to a human user. We illustrate several projections which have been found by artificial neural network extensions of Hebbian learning.

     We then show examples of similar projections found by reinforcement learning; our rationale in this case is that we have agents interacting proactively with a database examining different projections and, without human intervention, getting rewards when the projections reveal some interesting structure. We then give examples of the same projections found by other computational intelligence methods such as the cross entropy method and artificial immune systems.

     We then examine projections to nonlinear manifolds and show that with a particular model of an underlying latent space, we may identify clusters in data sets when such clusters are not visible in any low dimensional linear projection. We extend current methods by using Bregman divergences.

     Finally we review different data representation techniques: we begin with parallel coordinates and point out some difficulties with this method before reviewing Andrews’ Curves, a method from the 1970s which has only become truly practicable with the advent of modern desktop computers. An extension to this method came from Wegman and his colleagues in the 1990s. We also discuss a more recent extension and illustrate three dimensional projections of data samples dancing together.

Short Bio

     Professor Colin Fyfe is an active researcher in Artificial Neural Networks, Genetic Algorithms, Artificial Immune Systems and Artificial Life having written over 300 refereed papers, several book chapters and three books. He is a member of the Editorial Board of 5 journals. He currently supervises 3 PhD students and has acted as Director of Studies for 20 PhDs (all successful) since 1998. He has been Visiting Professor at universities in Spain, Australia, USA, Hong Kong, Taiwan and South Korea. He has been Honorary Chair at several recent international conferences, has given several plenary talks and tutorials.

Prof. Adam Grzech
Wroclaw University of Technology, Poland.

Title of talk

Intelligent distributed detection systems of computer communication systems

Abstract

    Continued growth of amount and complexity of services offered by providers and required by customers in contemporary distributed computer communication systems is resulting in an increasingly complex, interconnected infrastructure. In gain to assure efficiency, flexibility, quality and security of distributed communication systems, the infrastructure require intelligent management systems that are responsive, adaptive, proactive and less centralized than those deployed. Such required properties are offered by distributed approaches that give the potential to develop more advanced and effective network-based strategies replacing traditional node-based approaches.

    The talk is devoted to present various architectures of intelligent, distributed network-based intrusion detection systems and measures of distributed intrusion detection system quality. Moreover an impact of network and their intrusion detection system architectures parameters on the intrusion detection systems quality is discussed and illustrated.

Short Bio

    Adam Grzech received the M.S. degree in automatic control in 1977, Ph.D. degree in computer science in 1979 and D.Sc. degree in computer science from the Wroclaw University of Technology, Poland. Currently, he is a professor in the Institute of Computer Science, Wroclaw University of Technology. His research interests include computer communication networks, networks architecture and protocols, distributed communication systems, networks performance and quality of network services.

Prof. Kazumi Nakamatsu
University of Hyogo, Japan

Title of talk

Application of Paraconsistent Annotated Logic Program EVALPSN to Control/Safety Verification

Abstract

     I have already proposed a paraconsistent annotated logic program called Extended Vector Annotated Logic Program with Strong Negation (EVALPSN), which can deal with defensible deontic reasoning. EVALPSN has been applied to various intelligent control and safety verification systems such as pipeline valve control, railway interlocking safety verification, etc. Moreover, EVALPSN has been developed to deal with before-after relation between processes and it can be applied to process order control and its safety verification. The developed EVALPSN is called bf (before-after) EVALPSN. It will be introduced how to apply EVALPSN and bf-EVALPSN to intelligent control and safety verification with   examples and simulation results.

Short Bio

      Prof. Kazumi Nakamatsu has taken his Doctor of Science from Kyushu University 1999 and been a professor at School of Human Science and Environment, University of Hyogo since 2004. His research focuses on application of formal logics, especially paraconsistent annotated logic program, with applications to computer science area. He has developed a paraconsistent logic program called an EVALPSN, and applied it to intelligent control and safety verification for various systems such as railway interlocking safety verification, pipeline valve control, traffic signal control, etc. He has applied a PAT in terms of intelligent process order control based on bf-EVALPSN.