From 21 to 25 April 2018, a training workshop on the use of the GAMS software was held in Conakry, Guinea. The workshop focused on the main GAMS software commands, the intermediate commands, the advanced commands, the output files, their interpretation, the programming of a MEGC model and the stochastic simulations.
 
First, from an example of a program on fruit producers in Guinea, the trainer returned to the basic syntax used in GAMS. A practical exploration of the common features of GAMS programs, in this a practical example will be presented. It will be a question of minimizing the cost of transport of fruits in three capital of the sub-region. This example focuses on using the main building blocks of a GAMS program, such as declarations, definitions, and uses of SET, PARAMETER, TABLE, VARIABLE, EQUATIONS, MODEL, and SOLVE.

In a second time the training focused on the output files and the techniques of detection and correction of errors in Gams. This session focused on interpreting GAMS output files, including the ".lst" file. The same general linear programming model (LP) introduced in the previous part is used, but the model is executed with the introduction of several errors. Other files on other simple programs with errors have been executed. This allowed users to practice finding and correcting common errors encountered while running GAMS. About twenty programs in which errors are introduced have been executed.
In a third step, it was a question of introducing more complex syntaxes and to explore the types of VARIABLE and the constraints imposed on the VARIABLES. Users learn to define sets in different ways, to map different set elements, and to define bounds of VARIABLES. This part includes an optimization problem adapted to a problem developed by David Laborde of the International Food Policy Research Institute. The concept of running simulations is presented to users, who will be able to operate the model under different conditions.

Then the trainer presented complex syntaxes such as conditional expressions, repetitive instructions in the form of a loop (for, while, loop) as well as conditions. Import and export of Excel to GAMS files have been learned. The trainer then presented the mathematical expressions of a simplified computable general equilibrium model. The presentation emphasized the logic of computable general equilibrium modeling on GAMS. Then it was a question of programming in GAMS the simplified computable general equilibrium model presented above. Each equation will be translated into GAMS code. This part explained how to perform simulation with a MEGC. Several simulations were performed.
One session of the training explained how and why perform stochastic simulations with GAMS. She came back to the distribution laws in statistics and showed how to generate series with GAMS according to the law in question. A complete MEGC without going into technical details was presented. All the equations have been presented and explained. Another session was devoted to the presentation of Guinea's Social Accounting Matrix (SAM). The exhibition emphasized the logic of MCS and its utility in the construction of a MEGC. The training session focused on the full MEGC GAMS code built by the Economic Policy and Partnership (PEP) researchers network. The different phases of simulations on a complete MEGC have been reviewed.