Software products

EpODE

EPODE (ExPert system for Ordinary Differential Equations) is a numerical problem solving environment for initial values problems, i.e. a computer system that provides all the computational facilities necessary to solve the following target class of problems: y'(t)=f(y(t)), y(t0)=y0, where t in [t0,t0+T]. EPODE features include advanced solution methods, automatic and semiautomatic selection of solution methods, and ways to easily incorporate novel solution methods. Moreover, it uses the language of the above class of problems, so users can run them without specialized knowledge of the underlying computer hardware or software. It also exploits modern technologies such as interactive color graphics and networks of specialized services.

EPODE proposes the following facilities: friendly user interface for problem specification, automatic detection of the problem properties like linearity, separability, the greatest and the lowest eigenvalues of the Jacobian matrix at the initial value (symbolic computation of the system Jacobian), stiffness ratio, estimated time for the evaluation of the right side of the ODE system, friendly user interface for a difference method specification, automatic detection of the method properties like explicit/implicit schema, method accuracy, stability properties, one-step/ multistep, one-stage/ multistage, one-derivative/ multiderivative method, estimated time for solving the test problem y'=ey with Re e<0. It establish the matching between the problem properties and the method properties, adequate integration step, approximate computation time, estimated error, and then apply the numerical method and supervise the error. It is possible to create a list/graphic of approximate solution values (two- or three-dimensional graphs). It is also possible to solve the problem without the specification of a particular method: (i) choosing an appropriate method from a database of classical methods (the selection is based on a classification of the methods, and the database can be enlarged introducing new methods using the above mentioned front-end); (ii) select a new method when the previous one generates unreasonable errors; (iii) in the case of a large number of equations the expert recommends the use of the numerical codes combined with PVM procedures (Parallel Virtual Machine) in the idea to distribute the computations on some processors of a local network.The main thrust of EPODE is its automated identification of problem properties and method properties. The first version of EPODE was designed especially for the numerical solution of ordinary differential equations of stiff type (many software packages can not integrate such equations). EPODE is independent from any other computer application. Moreover, in EPODE, no supplementary code will be generated when a new method will be added.

Three actual versions of EPODE 1.0 are available for free download:

Further help concerning EpODE can be found here.

Remarks:

1. only the Windows '95 version has a context-help file at this moment;

2. in order to distribute the computations on a network of workstations you must have installed on your computer the PVM software;

3. any suggestion or remarks on the malfunction of the software can be addressed to the author .

PVMaple

PVMaple system proposes an extension of Maple capabilities to distributed computations for workstations grouped into a Parallel Virtual Machine. The user interacts with the system via the text oriented Maple front-end. PVMaple can be used in any network environment where Maple and PVM are available. It is similar with Distributed Maple (Java & Maple, available for Unix systems from http://www.risc.uni-linz.ac.at/software/distmaple).

Curent version of PVMaple 0.0 for Win9x and Unix available for free download includes command messenger master.exe, function definitions in Maple pvm.h, a .BAT file, an example of setting file in text format, short documentation and some examples

Instalation notes: PVMaple cannot function without PVM (Parallel Virtual Machine, http://netlib.org/pvm). The command messenger must be copied into the PVM directory with binaries, Maple path must be recognized, and pvm.h must be readed into a Maple document before any call in Maple of the form pvm[...].

Maple2g

Maple2g package allows the connection between Maple and computational grids based on the Globus Toolkit. The prototype of a grid-enabling wrapper for Maple, consists of two parts (a) a CAS-dependent and (b) a grid-dependent one.

  • m2g, the library of functions allowing the Maple user to interact with the grid/cluster middleware;
  • MGProxy, the middleware, a package of Java classes, acting as an interface between the m2g and the grid environment.
The m2g functions are implemented in the Maple language, and they call the MGProxy which accesses the Java CoG API. In this way, any change in the CAS or the grid will be reflected only in one part of the proposed system. The CAS-dependent part is relatively simple and can be easily modified to support another CAS or a legacy code.

Lastest codes can be found here.

DuctRot

DUCTROT (Ductility of Rotation) computer program provides the designer with the rotation capacity of steel members in order to verify the structure ductility.

The method of local plastic mechanism is used to determine the rotation capacity of steel members. This method has been proved to be the most effective one for practical design.

The DUCTROT M computer program is an interactive tool for computing rotation capacity of steel beams and beam-columns, elaborated as a result of many years of research works in this field by the research teams of Politechnica University Timisoara, Building Research Institute Timisoara and Western University Timisoara.

The program is designed to run on any PC with a version of Windows 95/98 or Millenium. Its small dimension (880Kb) make it easy to be moved from one workstation to another.

The graphical interface is based on a number of computational panels, each one having a specific meaning (cross-section types, material characteristics, etc.). A panel can have four types of fields: user input, results, buttons and explanation figures. Fields for data input (in white color) are provided for the user to describe his problem ( yield stress, flange width, axial force, etc.). Messages are supplied when the user gives some wrong data, i.e. out of normal range. The fields with results (in gray color) are filled with information provided by the program according the input data. Note that each modification of an input data automatically modifies the result values and also the effect of changing one data can be easily seen on those results which are on the same computational panel and are depending on that data. The buttons (with text or pictures on top) are used to select one from many choices or to navigate between panels (possible also using a scroll-bar).

The pictures can be classified into static and dynamic ones. Suggestive static pictures help the user to understand the notations used by the program. Note that each data field has associated a clear explanation on the bottom line (status line), which will appear when the user clicks in this field. If the user needs some supplementary information, he can activate a help window, where he can found proposed values for the respective characteristic or relationship used to determine the results. Dynamic pictures (diagrams) are those which are depending on the input data. A simple modification of the data on a figure panel or some previous data can affect the shape of the curves represented in that picture.

The program execution is user driven i.e. the computational panels depends on the data given by the user. So, according to the user inputs, different branches of the program can lead to different panels. In order to help the user to fill input data fields, each data has an implicit value which will be displayed when his resistant panel will be activated for the first time by the user. Any modification of a data field will influence the results on the current panel and those of the next panels. An order is established between panels: a panel do not appear until his previous one has not been activated.

The program menu help the user to save his current input (numerical values only) in a text file which can be used by another program or in another working session with DUCTROT. The implicit extension of the output file is "loc". The information to be printed can be select: the printed information can be a table with inputs or only those of specific panel results, dynamical diagrams (all or a specific one), or all information (inputs, numerical results in tables and diagrams). A toolbar allows the rapid access to the menu facilities (new session, open session, save session, print and exit program.

The program was written in Microsoft Visual C++ version 5.0. Therefore the program interface is specific for Windows applications, the open, save and print interfaces correspond to that of the specific Windows version on which the program is running. The diagrams are constructed using the Visual C tools for two-dimensional graphics.

A demo version of Windows operating systems is available for free download: