Multi-objective Estimation of Distribution Algorithm Based on Joint Modeling of Objectives and Variables

This paper proposes a new multi-objective estimation of distribution algorithm (EDA) based on joint modeling of objectives and variables. This EDA uses the multi-dimensional Bayesian network as its probabilistic model. In this way it can capture the dependencies between objectives, variables and objectives, as well as the dependencies learnt between variables in other Bayesian network-based EDAs. This model leads to a problem decomposition that helps the proposed algorithm to find better trade-off solutions to the multi-objective problem. In addition to Pareto set approximation, the algorithm is also able to estimate the structure of the multi-objective problem. To apply the algorithm to many-objective problems, the algorithm includes four different ranking methods proposed in the literature for this purpose. The algorithm is applied to the set of walking fish group (WFG) problems, and its optimization performance is compared with an evolutionary algorithm and another multi-objective EDA. The experimental results show that the proposed algorithm performs significantly better on many of the problems and for different objective space dimensions, and achieves comparable results on some compared with the other algorithms.

Objective optical classifier for tobacco leaves.

The main objective of the current research was to search the optimum method to segregate the most frequent color commercial quality classes of tobacco leaves (c.v. "Virginia"). These color classes cover the whole continuous color scale, between "Pale Lemon" and "Oxidated Brown". With the usual expert classification there exists a significant level of uncertainty . Within this research, several methods for data discrimination were tested, in order to solve uncertainty. Classification errors below 5% were obtained with this proposed classifier along two different seasons (1994&1995).

Laser induced forward transfer of silver pastes for printing of fingers in c-si cells.

The main objective of this work is to adapt the Laser Induced Forward Techniques (LIFT), a well- known laser direct writing technique for material transfer, to define metallic contacts (fingers and busbars) onto c-Si cells. The silver paste (with viscosity around 30-50 kcPs) is applied over a glass substrate using a coater. The thickness of the paste can be control changing the deposit parameters. The glass with the silver paste is set at a controlled gap over the c-Si cell. A solid state pulsed laser (532 nm) is focused at the glass/silver interface producing a droplet of silver that it is transferred to the c-Si cell. A scanner is used to print lines. The process parameters (silver paste thickness, gap and laser parameters -spot size, pulse energy and overlapping of pulses) are modified and the morphology of the lines is studied using confocal microscopy. Long lines are printed and the uniformity (in thickness and height) is studied. Some examples of metallization of larger areas (up to 10 cm x 10 cm) are presented.