Machine learning (ML) methods facilitate automated data mining. The authors compare the effectiveness of selected ML methods (RBF networks, Kohonen networks, and random forest) as modelling tools supporting the selection of materials in ecodesign. Applied in the design process, ML methods help benefit from the knowledge, experience and creativity of designers stored in historical data in databases. Implemented into a decision support system, the knowledge can be utilized – in the case under analysis – in the process of design of environmentally friendly products. The study was initiated with an analysis of input data for the selection of materials. The input data, specified in cooperation with designers, include both technological and environmental parameters which guarantee the desired compatibility of materials. Next, models were developed using selected ML methods. The models were assessed and implemented into an expert system. The authors show which models best fit their purpose and why. Models supporting the selection of materials, connections and disassembly methods help boost the recycling properties of designed products.
This article presents a computer system for the identification of casting defects using the methodology of Case-Based Reasoning. The
system is a decision support tool in the diagnosis of defects in castings and is designed for small and medium-sized plants, where it is not
possible to take advantage of multi-criteria data. Without access to complete process data, the diagnosis of casting defects requires the use
of methods which process the information based on the experience and observations of a technologist responsible for the inspection of
ready castings. The problem, known and studied for a long time, was decided to be solved with a computer system using a CBR (CaseBased
Reasoning) methodology. The CBR methodology not only allows using expert knowledge accumulated in the implementation
phase, but also provides the system with an opportunity to "learn" by collecting new cases solved earlier by this system. The authors
present a solution to the system of inference based on the accumulated cases, in which the main principle of operation is searching for
similarities between the cases observed and cases stored in the knowledge base.
This article presents a practical solution in the form of implementation of agent-based platform for the management of contracts in
a network of foundries. The described implementation is a continuation of earlier scientific work in the field of design and theoretical
system specification for cooperating companies [1]. The implementation addresses key design assumptions - the system is implemented
using multi-agent technology, which offers the possibility of decentralisation and distributed processing of specified contracts and tenders.
The implemented system enables the joint management of orders for a network of small and medium-sized metallurgical plants, while
providing them with greater competitiveness and the ability to carry out large procurements. The article presents the functional aspects of
the system - the user interface and the principle of operation of individual agents that represent businesses seeking potential suppliers or
recipients of services and products. Additionally, the system is equipped with a bi-directional agent translating standards based on
ontologies, which aims to automate the decision-making process during tender specifications as a response to the request.
The main scope of the article is the development of a computer system, which should give advices at problem of cooper alloys
manufacturing. This problem relates with choosing of an appropriate type of bronze (e.g. the BA 1044 bronze) with possible modification
(e.g. calcium carbide modifications: Ca + C or CaC2) and possible heat treatment operations (quenching, tempering) in order to obtain
desired mechanical properties of manufactured material described by tensile strength - Rm, yield strength - Rp0.2 and elongation - A5. By
construction of the computer system being the goal of presented here work Case-based Reasoning is proposed to be used. Case-based
Reasoning is the methodology within Artificial Intelligence techniques, which enables solving new problems basing on experiences that
are solutions obtained in the past. Case-based Reasoning also enables incremental learning, because every new experience is retained each
time in order to be available for future processes of problem solving. Proposed by the developed system solution can be used by
a technologist as a rough solution for cooper alloys manufacturing problem, which requires further tests in order to confirm it correctness.
In flowering plants, seeds are produced both sexually (double fertilization is required) and asexually via apomixis (meiotic reduction and egg fertilization are omitted). An apomictic-like pattern of endosperm development in planta is followed by fis mutants of sexual Arabidopsis thaliana. In our experiments in planta, autonomous endosperm (AE) developed in met1 mutants. Furthermore we obtained autonomous endosperm formation in vitro not only in unfertilized ovules of fie mutants but also in wild genotypes (Col-0, MET1/MET1, FIE/FIE) and met1 mutants. AE induction and development occurred in all genotypes on the each of the media used and in every trial. The frequency of AE was relatively high (51.2% ovaries) and genotype-dependent. AE induced in vitro represents a more advanced stage of development than AE induced in fie mutants in planta. This was manifested by a high number of nuclei surrounded by cytoplasm and organized in nuclear cytoplasmic domains (NCDs), nodule formation, division into characteristic regions, and cellularization. The high frequency of AE observed in homozygous met1 (met1/met1) mutants probably is due to accumulation of hypomethylation as an effect of the met1 mutation and the in vitro conditions. AE development was most advanced in FIE/fie mutants. We suggest that changes in the methylation of one or several genes in the DNA of Arabidopsis genotypes caused by in vitro conditions resulted in AE induction and/or further AE development.
The binary classifiers are appropriate for classification problems with two class labels. For multi-class problems, decomposition techniques, like one-vs-one strategy, are used because they allow the use of binary classifiers. The ensemble selection, on the other hand, is one of the most studied topics in multiple classifier systems because a selected subset of base classifiers may perform better than the whole set of base classifiers. Thus, we propose a novel concept of the dynamic ensemble selection based on values of the score function used in the one-vs-one decomposition scheme. The proposed algorithm has been verified on a real dataset regarding the classification of cutting tools. The proposed approach is compared with the static ensemble selection method based on the integration of base classifiers in geometric space, which also uses the one-vs-one decomposition scheme. In addition, other base classification algorithms are used to compare results in the conducted experiments. The obtained results demonstrate the effectiveness of our approach.
The aim of this study is to design and implement a computer system, which will allow the semantic cataloging and data retrieval in the
field of cast iron processing. The intention is to let the system architecture allow for consideration of data on various processing techniques
based on the information available or searched by a potential user. This is achieved by separating the system code from the knowledge of
the processing operations or from the chemical composition of the material being processed. This is made possible by the creation and
subsequent use of formal knowledge representation in the form of ontology. So, any use of the system is associated with the use of
ontologies, either as an aid for the cataloging of new data, or as an indication of restrictions imposed on the data which draw user attention.
The use of formal knowledge representation also allows consideration of semantic meaning, a consequence of which may be, for example,
returning all elements in subclasses of the searched process class or material grade.
The problem of materials selection in terms of their mechanical properties during the design of new products is a key issue of design. The
complexity of this process is mainly due to a multitude of variants in the previously produced materials and the possibility of their further
processing improving the properties. In everyday practice, the problem is solved basing on expert or designer knowledge. The paper is the
proposition of a solution using computer-aided analysis of material experimental data, which may be acquired from external data sources.
In both cases, taking into account the rapid growth of data, additional tools become increasingly important, mainly those which offer
support for adding, viewing, and simple comparison of different experiments. In this paper, the use of formal knowledge representation in
the form of an ontology is proposed as a bridge between physical repositories of data in the form of files and user queries, which are
usually formulated in natural language. The number and the sophisticated internal structure of attributes or parameters that could be the
criteria of the search for the user are an important issue in the traditional data search tools. Ontology, as a formal representation of
knowledge, enables taking into account the known relationships between concepts in the field of cast iron, materials used and processing
techniques. This allows the user to receive support by searching the results of experiments that relate to a specific material or processing
treatment. Automatic presentation of the results which relate to similar materials or similar processing treatments is also possible, which
should make the conducted analysis of the selection of materials or processing treatments more comprehensive by including a wider range
of possible solutions.
The objective of studies presented in this publication was structuring of research knowledge about the ADI functional properties and
changes in these properties due to material treatment. The results obtained were an outcome of research on the selection of a format of
knowledge representation that would be useful in further work aiming at the design, application and implementation of an effective system
supporting the decisions of a technologist concerning the choice of a suitable material (ADI in this case) and appropriate treatment process
(if necessary). ALSV(FD) logic allows easy modelling of knowledge, which should let addressees of the target system carry out
knowledge modelling by themselves. The expressiveness of ALSV (FD) logic allows recording the values of attributes from the scope of
the modelled domain regarding ADI, which is undoubtedly an advantage in the context of further use of the logic. Yet, although the logic
by itself does not allow creating the rules of knowledge, it may form a basis for the XTT format that is rule-based notation. The difficulty
in the use of XTT format for knowledge modelling is acceptable, but formalism is not suitable for the discovery of rules, and therefore the
knowledge of technologist is required to determine the impact of process parameters on values that are functional properties of ADI. The
characteristics of ALSV(FD) logic and XTT formalism, described in this article, cover the most important aspects of a broadly discussed,
full evaluation of the applicability of these solutions in the construction of a system supporting the decisions of a technologist.
One way to ensure the required technical characteristics of castings is the strict control of production parameters affecting the quality of
the finished products. If the production process is improperly configured, the resulting defects in castings lead to huge losses. Therefore,
from the point of view of economics, it is advisable to use the methods of computational intelligence in the field of quality assurance and
adjustment of parameters of future production. At the same time, the development of knowledge in the field of metallurgy, aimed to raise
the technical level and efficiency of the manufacture of foundry products, should be followed by the development of information systems
to support production processes in order to improve their effectiveness and compliance with the increasingly more stringent requirements
of ergonomics, occupational safety, environmental protection and quality. This article is a presentation of artificial intelligence methods
used in practical applications related to quality assurance. The problem of control of the production process involves the use of tools such
as the induction of decision trees, fuzzy logic, rough set theory, artificial neural networks or case-based reasoning.
This article presents the methodology for exploratory analysis of data from microstructural studies of compacted graphite iron to gain
knowledge about the factors favouring the formation of ausferrite. The studies led to the development of rules to evaluate the content of
ausferrite based on the chemical composition. Data mining methods have been used to generate regression models such as boosted trees,
random forest, and piecewise regression models. The development of a stepwise regression modelling process on the iteratively limited
sets enabled, on the one hand, the improvement of forecasting precision and, on the other, acquisition of deeper knowledge about the
ausferrite formation. Repeated examination of the significance of the effect of various factors in different regression models has allowed
identification of the most important variables influencing the ausferrite content in different ranges of the parameters variability.