Frequently Asked Questions

1.- What do I need to use SSm GO toolbox?
A PC with Windows XP/Vista/7 (32 or 64 bits), Matlab 7.0-7.8 (32 bits version), and optionally the Mathworks Optimization Toolbox

1b.- I have a 64 bits version of Matlab, can I use SSm GO?
No, you need a 32 bits version of Matlab. See reply 1d below for a solution.

1c.- I have a 64 bits version of Windows, is this a problem?
No as long as you have a 32 bits version of Matlab

1d.- I have Matlab R2010a and the toolbox does not work, why? How can I solve this?
It seems that mex files are not backwards compatible in the latest versions of Matlab.
We can not invest more time solving these issues for every new Matlab version, so we advice you to install an older (e.g. 7.5 - 7.8) Matlab version in addition to your existing one.
Note that the Mathworks allows you to do this (you can download a previous version from their web site as long as you have a valid license).

2.- Do you have a Linux version? A version for Mac?
Linux: several of the local solvers of SSm GO are mex files for Windows and therefore do not run under Linux. Porting them to Linux has proved a non-trivial exercise, so currently we do not offer a Linux version. However, you can still use the toolbox if you avoid calling these local solvers (note, however, that in many cases this will result in worse results and efficiency)
Mac: see Linux reply above

3.- Where can I get the toolbox?
Go to http://www.iim.csic.es/~gingproc/ssmGO.html and follow the instructions

4.- I work in a company and I would like to use this toolbox for commercial purposes, is it possible?
No, currently we only provide the toolbox to academic, non-profit users.

5.- What is the license type of SSm GO?
Please check the bottom of http://www.iim.csic.es/~gingproc/ssmGO.html

6.- How many global optimization solvers does the toolbox include?
Currently, there are two, eSS and SSm. Both are based on scatter search, but we highly recommend you to use eSS. We include SSm to ensure backwards compatibility with previous versions of the toolbox

7.- Where can I find help and documentation? And examples?
Go to the documentation page at http://ssmgot.posterous.com/pages/documentation
For quick help, check http://ssmgot.posterous.com/pages/quick-help-ssm-and-ess-solvers
(note: these documents are also included in the software distribution)
For examples, check the User's Guide in the documentation link above

8.- I have obtained some results and I am going to publish them. How should I cite SSm GO?
If you have used the eSS solver, cite:
Egea, J.A., Martí, R., & Banga, J.R. (2010). An evolutionary method for complex-process optimization. Computers and Operations Research 37(2):315-324.
If you have used the SSm solver, cite:
Egea, J.A., M. Rodriguez-Fernandez, J. R. Banga and R. Martí (2007) Scatter Search for chemical and bioprocess optimization. Journal of Global Optimization 37(3):481-503.
If you have used the toolbox for parameter estimation, you can also cite:
Rodriguez-Fernandez, M., J. A. Egea and J. R. Banga (2006) Novel Metaheuristic for Parameter Estimation in Nonlinear Dynamic Biological Systems. BMC Bioinformatics 7:483

9.- Have you compared the solvers in SSm GO with other global optimizers, like genetic algorithms or simulated annealing?
Yes, check the comparison at http://ssmgot.posterous.com/ssm-go-comparison-with-other-optimization-met

10.- I want to start using SSm GO, but I have many doubts regarding the tuning of its parameters. Please help me!
Please check the User's Guide (see 7 above) and the examples. Default values for the search parameters seem to be OK for most problems, but you can try changing them following the User's Guide. We will soon provide some quick hints for tuning the methods here, stay tuned.

11.- I have tried to use your toolbox with my problem, but I get errors, please help.
Please check first the User's Guide and examples to make sure you are calling the solvers in the correct way.
If you still get errors, please send an email to gingproc@iim.csic.es with the details and the scripts for your problem (it is difficult for us to help you if we can not reproduce the error). If you want to maintain your problem scripts secret, convert them to p-code and then send that to us.

12.- Can the methods in SSm GO guarantee the convergence to global solutions?

No, these methods are stochastic, so no guarantees. You can only get such guarantees with certain types of deterministic global optimization methods, BUT currently these methods can only solve small problems in reasonable computation times. Plus, our solvers in SSm GO can be applied to problems stated as black-boxes (they only need function evaluation values, no need of gradients or any other info), something which is not possible for most deterministic methods. Besides, empirically, the solvers at SSm GO have been able to solve challenging global optimization problems, so although they do not offer guarantees, in practice they seem to work quite or even very well for many applications. Check them with your own problem!