New Features AIMMS 3.11

Home

Production and Feature Releases

In our AIMMS release cycle we release Production and Feature Releases. A series of Feature Releases is followed by its Production Release. The most current Production Release is AIMMS 3.11 (March 2011).

In May 2011 we have started the release of our Feature Release series of AIMMS 3.12. Our current release (as of December 2011) AIMMS 3.12 Feature Release 3 is available for download.  

AIMMS 3.11 Production Release

On this page you can read about the new features in AIMMS 3.11. More details can be found in the release notes. Please note that we have added the CBC and IPOPT solvers to the AIMMS 3.11 Production Release.

Features introduced in AIMMS 3.11 FR1 (May 2010)

Features introduced in AIMMS 3.11 FR2 (September 2010)

Features introduced in AIMMS 3.11 FR3 (December 2010)

New Features AIMMS 3.11

Below, you can find a description of all new features that have been introduced in AIMMS 3.11. The Feature Release in which the features have first been introduced, are mentioned after the feature name, in brackets, where FRx means Feature Release x.

CBC and IPOPT solvers

Since the initial release of the solver link to COIN-OR solvers CBC and IPOPT in March 2010, we have extended the platform availability of this link to cover use of the solvers for all 32/64-bit Windows and Linux AIMMS versions. In addition, we have added the CBC and IPOPT solvers into the distribution files of AIMMS 3.11 to minimize the hurdle to use these open source solvers. The AIMMS link and the solvers remain available at COIN-OR and are released under the Eclipse Public License 1.0. 

Robust Optimization add-on (FR1)

As we see more and more customers trying to deal with the uncertainty of data (instead of focusing on single point solutions), it was only natural for us to work with Professor A. Ben-Tal , developer of the Robust Optimization methodology together with Professor A. Nemirovski. Robust Optimization (short RO) is an uncertainty modeling approach suitable for a situation where the range of the uncertainty is known and not necessarily the distribution (depending on the industry, this could be price, temperature, demand, etc). Typically some inputs take an uncertain value anywhere between a fixed minimum and maximum. Solutions will be feasible for all the constraints when the inputs drift within the uncertainty ranges. If this is too strict, one can even provide a probability for which the solution is required to satisfy the constraints with at least this probability (e.g., the chance that the demand for electricity will be met is at least 95%).

As of 3.11 FR1, AIMMS offers support for generating a Robust Optimization model from any given deterministic LP model, without the need to reformulate any of the constraint definitions. By only supplying additional attributes for selected parameters, variables and constraints, AIMMS can generate both a deterministic and robust model from the same formulation. A deterministic model, and a robust optimization model can therefore co-exist within the same master model and their respective solutions can be compared.

Availability: The RO add-on is available under AIMMS 3.11. Commercial users can extend their licenses with Robust Optimization, by purchasing the RO add-on. In general, any LP solver can be used for solving RO models. However, for a specific class (i.e. when the uncertainty is defined as an ellipsoid), CPLEX or MOSEK are required. The RO add-on itself is included at no charge in the AIMMS Educational licenses.

Stochastic Benders speed-up (FR1)

After our first implementation of the Stochastic Benders algorithm (released in AIMMS 3.8, Dec 2007), we learned that the performance should be improved to make it usable for a larger group of customers (especially the ones with large models, or many scenarios). We therefore decided to do a partly rewrite of the underlying code that resulted in a speed up of 1.5 to 40 times on our test set. Thanks to the rewrite, we have also been able to improve the memory usage drastically. This new implementation of Stochastic Benders will enable you to solve your stochastic models faster, or solve bigger instances of your problem (which become impractical due to timing issues, or simply ran out of memory).

Connection Strings for database identifiers (FR1)

When making a database connection from AIMMS, users either had to use a system/user data source link, or a .dsn/.udl file. The first method has the disadvantage that they are machine dependent, making transfers of models not straightforward; the latter method is machine independent, but when combining it with username/passwords (which is commonplace in the database world), passwords would be stored in readable format in the .dsn/.udl files. To be able to allow easy transfer between machines and hide username /passwords, we have implemented the concept of connection strings in AIMMS. Connection strings can be used in the data source attribute of database related identifiers; they typically contain all relevant connection information, including username/password combinations, and can dynamically be created from within your model.

Network License improvements (FR1)

To increase the usability and availability of AIMMS licenses that are shared through the AIMMS Network License Manager to a group of AIMMS users and to make management of the licenses by IT easier, we have added the following features:

  • Support for multiple default profiles. This means that if a user requests a license without a specified profile, the first available default profile that is not fully in use will be provided (instead of a ‘no availability’ result) .
  • If AIMMS is run on Windows and fails to connect to a License Server, users of AIMMS Developer or AIMMS End User systems will see a connect dialog box which shows that AIMMS repeatedly tries to connect. If required, the user can stop these attempts through this dialog.
  • A new command line argument allows you to specify a connection timeout interval. If all network licenses are in use, the logon will wait for the specified time until a license becomes available.
    Requirement: Installation of the latest AIMMS Network License Manager.

Hiding Composite Table columns by end user (FR1)

As our customers (in fact the end users of AIMMS applications) asked for more control in the Composite Tables on what columns to see and in what order, the developer can now define these controls for their end users. If needed, the changes can be saved. This change benefits both end users and developers of AIMMS applications as end users gain more control and developers don’t need to be involved in these changes once it has been set up.

Improved AIMMS Google group support (FR1)

Since the AIMMS Google Group (our AIMMS forum) has been successfully active for more than a year now, we want to make it easier for AIMMS users to get insight and interact in the discussions. We therefore include the most recent posts directly on the AIMMS Start Page (next to the online documentation, latest news etc.). Hyperlinks will bring you directly to the full thread through your web browser after which you can continue to browse the AIMMS Google Group. So next time you start up AIMMS, check out the latest posts, or go directly to opent in een nieuw venster the AIMMS Google group.

Other items (FR1)

In addition to the highlighted features, with AIMMS 3.11 FR1 we offer extended/new features such as Windows Server 2008 support, extended solver configuration support, software nodelock license for
educational users, two Robust Optimization examples, and a Gate Assignment example.

Model Edit functions for dynamic model additions (FR2)

Based on our customer request to offer the ability to dynamically add variables and constraints, or a custom objective to an AIMMS model, we are now offering a brand new set of Model Edit functions. Although initially designed to allow AIMMS modelers to include code that gives their end-users the ability to customize mathematical models on the fly, we have extended the design to offer much more.

The brand new Model Edit functions of AIMMS, allow you to dynamically create, execute, use, and modify model code (identifiers, procedures and such). This means you can perform tasks and create features in your AIMMS solutions that were not possible before. Such as automated modeling tasks (e.g. database interface generation; see example with this name), handling end-user proprietary information (e.g. input blend rules), or creating dynamic data reports (e.g. specify custom query over identifiers). To support the Model Edit functions, several Model Query functions have been added.

The Model Edit functions are made available from within AIMMS and through the AIMMS API. For details on this functionality, check our documentation and examples:

  • Language Reference (chapter The AIMMS Programming Interface, section Model Editing, and chapter Model Structure and Models, section Runtime Libraries and the Model Edit Functions),
  • Function Reference, (chapter Model Query Functions and chapter Model Edit Functions, and
  • Database Interface Generation example and Model Edit Functions example.

If you want to learn more, please contact our support group.

Robust Optimization add-on for MIP models (FR2)

After successfully introducing the Robust Optimization add-on in AIMMS 3.11 FR1 to support uncertainty modeling for LP’s, the Robust Optimization add-on can now also handle MIP models. In addition to the MIP extension, AIMMS now - by default - transforms ellipsoidal uncertainty constraints (both for LP and MIP), resulting in a more numerical stable robust counterpart. For details on all RO extensions, see the release notes.

Availability: The RO add-on is available as of AIMMS 3.11 FR1. Commercial users can extend their licenses with Robust Optimization, by purchasing the RO add-on. In the AIMMS Educational licenses the RO add-on is included at no charge.

Updated solvers (FR2)

AIMMS 3.11 FR2 offers CPLEX 12.2 and GUROBI 3.0.

Automatic resizing in Pivot Table (FR2)

The Pivot Table has been extended with a new option to automatically resize the column widths upon data changes. This option can be found on the Column Area tab of the Pivot Table properties. 

PageGetAll (FR2)

With the new function PageGetAll you can easily fill a set with the names of all pages and/or templates in your project. With the additional optional arguments, you can decide to exclude pages that are currently hidden, or pages and templates that are only used for printing.

Geographical coordinates (FR2)

After some time of absence, the AIMMS function to collect the geographical coordinates of addresses has been reintroduced (GeoFindCoordinates). Ideal if you work with GIS data which then also can be displayed in the AIMMS Network Object as it has GIS support.

Insert/Delete functionality for the Pivot Table (FR3)

In the Pivot Table, you can now easily insert a row or column for those elements for which there is currently no data available. You simply press the Insert key and AIMMS will let you select an element that is currently not visible. After that you can edit the data for this element.

Similarly, with the Delete key you can easily remove an entire row or column from the current table. Note that the data for the removed element is not deleted: the row or column is only deleted from the current view.

Index Filters in the Pivot Table (FR3)

The Pivot Table has been extended with Index Filters that allow you to view only a subset of the available data. The several types of filters make it possible to filter on either a number of explicitly selected elements, or on elements whose name match a simple search expression.

You can also easily switch from sparse to dense data display via the Index Filter pop-up menu.

New solver versions: GUROBI 4.0 and KNITRO 7.0 (FR3)

GUROBI 4.0 supports models with quadratic objective functions. The new version includes primal simplex, dual simplex, and parallel barrier optimizers for continuous QP models, and a parallel branch-and-cut solver for Mixed Integer Quadratic Programming (MIQP) models. The MIP solver is faster in release 4.0. The simplex and barrier solvers are slightly faster in release 4.0. GUROBI 4.0 also improves the numerical stability of the primal simplex solver and the barrier crossover algorithm.

KNITRO 7.0 introduces several improvements in linear algebra to improve performance.

New function FileGetSize (FR3)

The function FileGetSize has been added to AIMMS. With this function, you can retrieve the size of a file.

Model Edit functions: runtime libraries can be saved in cases (FR3)

As an addition to the Model Edit Functions introduced in AIMMS 3.11 FR2, runtime libraries (the model editor structure, the identifiers themselves and their data) can now be saved into cases. The option 'Case contains runtime libraries' has been introduced to enable saving.

Model Edit API example (FR3)

A new example has also been added to AIMMS (Model Edit API). It illustrates the use of the Model Edit functions from the AIMMS API by creating a simple transport model.
 

Previous releases contain the following new features:

 


 

 Request Free Trial  Discover the benefits of AIMMS

 Newsletters  Read our latest newsletter

 Login  to update your profile.

Login