26.11.2012 Views

Technical Sessions – Monday July 11

Technical Sessions – Monday July 11

Technical Sessions – Monday July 11

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

4 - Development of Ubiquitous Sensing System to Visualize<br />

Current Status of Production of Plant Manufacturing<br />

Industries<br />

Kazuyoshi Tsurusaki, Faculty of Economics, Nagasaki<br />

University, 4-2-1 Katafuchi, 850-8506, Nagasaki, Japan,<br />

turusaki@nagasaki-u.ac.jp<br />

The product management, which has been evolved continuously in massproduction<br />

goods, such as cars and electronic products since the beginning<br />

of last century, has been left behind in the evolution in plant industries such<br />

as power plants. One of key issues to prevent the evolution is considered to<br />

be its difficulty of visualization of production status in plant manufacturing.<br />

The authors examined several procedures to sense status of manufacturing with<br />

ubiquitous sensors attached on equipments and materials in production in the<br />

factories and demonstrated it at a factory.<br />

� TA-13<br />

Tuesday, 9:00-10:30<br />

Meeting Room 206<br />

Perturbations, Graphs and Games<br />

Stream: Continuous and Non-Smooth Optimization<br />

Invited session<br />

Chair: Jerzy Filar, Mathematics and Statistics, University of South<br />

Australia, Mawson Lakes Blvd, 5095, Mawson Lakes, SA, Australia,<br />

j.filar@unisa.edu.au<br />

Chair: Vladimir Gaitsgory, University of South Australia, 5095,<br />

Mawson Lakes, South Australia, Australia,<br />

Vladimir.Gaitsgory@unisa.edu.au<br />

1 - Polynomial Limit Control Algorithm to Identify Nearly all<br />

Cubic, Non-Hamiltonian, Graphs<br />

Jerzy Filar, Mathematics and Statistics, University of South<br />

Australia, Mawson Lakes Blvd, 5095, Mawson Lakes, SA,<br />

Australia, j.filar@unisa.edu.au, Ali Eshragh Jahromi<br />

Determining whether a cubic graph is non-Hamiltonian is an NP-complete<br />

problem. However, by embedding the Hamiltonian Cycle Problem in the space<br />

of occupational measures of a discounted Markov control problem, we can construct<br />

a limit-control linear program that correctly identifies a majority of non-<br />

Hamiltonian cubic graphs. Furthermore, by adding graph specific constraints<br />

to the latter, all but extremely rare instances can be correctly identified.<br />

2 - Non-Linearity in Non-Zero Sum Games with Incompetent<br />

Players<br />

Justin Beck, Centre for Industrial and Applied Mathematics,<br />

University of South Australia, 5095, Mawson Lakes, South<br />

Australia, Australia, justin.d.beck@hotmail.com<br />

Game Theory is a well established area of mathematics dealing with problems<br />

where there is more than one player. In these "games’ the fortunes of the players<br />

are coupled by their actions. An implicit assumption in many of these models<br />

is the players are capable of executing their optimal strategies. However, in<br />

reality a player may not have the ability to execute their chosen strategy. This<br />

presentation will explore this topic with reference to theory and applications<br />

with a focus on the non-linear and discontinuous behaviour which can arise in<br />

non-zero sum games.<br />

3 - Genetic Theory of Cubic Graphs<br />

Michael Haythorpe, School of Mathematics and Statistics,<br />

University of South Australia, Mawson Lakes Boulevard,<br />

Mawson Lakes, 5095, Adelaide, SA, Australia,<br />

michael.haythorpe@unisa.edu.au, Pouya Baniasadi, Vladimir<br />

Ejov, Jerzy Filar<br />

Cubic graphs are widely researched, as they represent arguably the simplest<br />

subset of graphs for which common graph theory problems remain non-trivial.<br />

We investigate the generation of cubic graphs, and show that they can (in polynomial<br />

time) be identified as one of two types - genes and descendents. We<br />

show that any descendent can be generated from a set of genes via the use<br />

of six breeding operations. We consider the application of this theory to the<br />

Hamiltonian cycle problem, whereby finding a Hamiltonian cycle (HC) in a<br />

graph is reduced to finding the HCs of that graph’s ancestor genes.<br />

IFORS 20<strong>11</strong> - Melbourne TA-14<br />

4 - Pseudo Singularly Perturbed Linear Programs<br />

Vladimir Gaitsgory, University of South Australia, 5095,<br />

Mawson Lakes, South Australia, Australia,<br />

Vladimir.Gaitsgory@unisa.edu.au<br />

We study a linear programming problem with a linear perturbation introduced<br />

through a small parameter epsilon. We identify and analyze an unusual asymptotic<br />

phenomenon. Namely, discontinuous limiting behavior of the optimal<br />

objective function value of such a linear program may occur even when the<br />

rank of the coefficient matrix of the constraints is unchanged by the perturbation.<br />

Under mild conditions, this phenomenon is a result of the classical Slater<br />

constraint qualification being violated at the limit. An iterative, constraint augmentation<br />

approach for resolving this problem is proposed.<br />

� TA-14<br />

Tuesday, 9:00-10:30<br />

Meeting Room 207<br />

Global Optimization<br />

Stream: Continuous and Non-Smooth Optimization<br />

Invited session<br />

Chair: Nelson Maculan Filho, COPPE / PESC, Universidade Federal<br />

do Rio de Janeiro, Rio de Janeiro, RJ, Brazil, maculan@cos.ufrj.br<br />

Chair: Emilio Carrizosa, Estadistica e Investigacion Operativa,<br />

Universidad de Sevilla, Matematicas, Reina Mercedes s/n, 41012,<br />

Sevilla, Spain, ecarrizosa@us.es<br />

Chair: Rommel Regis, Mathematics, Saint Joseph’s University, 5600<br />

City Avenue, 19131, Philadelphia, PA, United States, rregis@sju.edu<br />

1 - Effective and Efficient Hybrid Methods for Solving<br />

Global Optimization Problems Including the Lennard-<br />

Jones Potential Energy Global Optimization Problem<br />

Jiapu Zhang, CIAO & School of ITMS, The University of<br />

Ballarat, MT Helen Campus Ballarat University, 3350, Ballarat,<br />

VIC, Australia, j.zhang@ballarat.edu.au<br />

In recent years large-scale global optimization (GO) problems have drawn considerable<br />

attention. These problems have many applications, in particular in<br />

data mining and biochemistry. Some successful hybrid methods and numberical<br />

computing experiences for GO will be reported.<br />

The Lennard-Jones potential energy minimization problem is a benchmark for<br />

testing new GO algorithms. It is studied through the optimal atomic-resolution<br />

molecular structure constructions of amyloid fibrils. This is very useful in furthering<br />

the goals of medicinal chemistry.<br />

2 - Optimal Controlled-limit and Preventive Maintenance<br />

Policy within Life Cycle of Products<br />

Wen Liang Chang, Department of Information Management,<br />

Cardinal Tien College of Healthcare & Management, Taipei,<br />

Taiwan, D9101402@mail.ntust.edu.tw<br />

This paper investigates controlled-limit and preventive maintenance (PM) policy.<br />

When the product fails, the failed product is rectified using minimal repair.<br />

In order to reduce the number of product failures, the seller performs imperfect<br />

PM actions when the age of the product reaches a controlled-limit. After<br />

the base warranty period expires, the seller gives a discount of purchasing extended<br />

warranty (EW) expense if the consumer purchases the EW for products.<br />

Under this maintenance scheme, the profit model is constructed and then the<br />

optimal policy is obtained. Finally, numerical examples are given to illustrate<br />

the influences of the optimal policy for profit model.<br />

3 - A Continuous Optimization Approach for Determining<br />

the Stability Number of a Graph<br />

Jitamitra Desai, School of Mechanical and Aerospace<br />

Engineering, Nanyang Technological University, Systems and<br />

Engineering Management, 50 Nanyang Avenue, N3.2 - 02-53,<br />

639798, Singapore, jdesai@ntu.edu.sg<br />

In this talk, we determine the stability number of a graph via a fractional programming<br />

formulation. The problem of finding the stability number is a wellstudied<br />

problem in IP literature, but solving this problem using continuous optimization<br />

techniques is a relatively new area of research. We showcase several<br />

convexity results, and prove that the fractional program yields the stability<br />

number. Characterizations of the local maxima are described, and these results<br />

are utilized in developing a new optimization algorithm. Detailed computational<br />

results are provided using standard test sets.<br />

35

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!