Stephanie,
nous avons complété le compte-rendu avec des questions et une petite
projection technique sur les échanges MSP/ISIS
On Fri, 24 May 2019 17:34:31 +0200
Stephanie MAHEVAS <Stephanie.Mahevas(a)ifremer.fr> wrote:
> Meeting report
>
> May 6th 2019, by phone
>
> Present : Mahevas Stephanie (Ifremer), Vincent Badts (Ifremer),
> Harald Warmelink (Breda University, Warmelink.H(a)buas.nl) Carlos
> (Breda University), Jean Couteau (Code Lutin, couteau(a)codelutin.com)
> Eric Chatellier (Code Lutin, chatellier(a)codelutin.com)
>
> Subject : Project Mimi, exchanges between computer scientists to
> consider connecting MSP to ISISFISH
>
> Report :
>
> * the time step of the exchanges between MSP and ISIS should be the
> month in order to favor the interaction of the players with the
> simulator
> * the simulation time must be limited (a few seconds max) for the
> interest of the game
> * the player can modify parameters during the simulation. It is
> therefore necessary to be able to return new parameters to
> ISISFISH, to stop the simulation, to configure ISISFISH and to
> restart the simulation
> * ...
>
>
> To be done :
>
> * list the maritime activities that we want to visualize on the game
> (Pierre Issac email, waiting for his reply)
> * specify the area of the polygons in MSP and in ISISFISH (probably
> coarser in ISIFISH) and build the transfer mode to MSP : in
> ISIS-Fish a regular grid
> * analyze the potential need for an application between MSP and
> ISISFISH to manage the flows between the two applications
> * provide a sufficiently powerful server machine to run ISISFISH and
> MSP in game mode
> * list the ISISFISH business rules that will force the player into
> the MSP interface
> * ...
>
> Next meeting : ....
>
> Scenarios that could be implemented step by step (increasing the
> complexity) 1. ISIS runs independantly and provides each month some
> outputs that can be shown dynamically in MSP
ISIS can already be launched with a prescript and can output results
each time step (file is updated after each time step). This scenario
could be very straight forward (ISIS speaking).
> 2. at t=0, some management variables or zones dedicated to other
> activities than fishing (that impact fishing activities in ISIS-Fish)
> are set in MSP, ISIS is fed with these variables, and then
> simulation as in scenario 1
This step could already be quite straight forward as the variables are
in the prescript as in step 1.
> 3. sceanario 2, and at any simulation time, some variables (as in
> scenario2) are set again, ISIS is updated with this new information
> and re-starts the simulation
For this step we need to think about the communication process between
ISIS and MSP.
>
> + example of ISIS-Fish outputs and prescripts
Technical Questions :
1. Is MSP able to launch an external process with parameters (eg.
launch ISIS - java process - with a prescript file as parameter) and
read result csv files ? If yes, step 1 and 2 should be very straight
forward regarding ISIS-Fish
2. For the dialog between MSP and ISIS, can we use an exchange file,
written by MSP and read by ISIS-Fish ? If yes, the communication/run
process could be :
1. Parameterize and launch run on MSP
2. Step 1 run in ISIS Fish with results produced
3. Results are read by MSP and printed to the user
4. ISIS waits for exchange file to be written by MSP
5. MSP asks the users if they want to make modifications if yes, records
the modifications
6. MSP writes to the exchange file
7. ISIS reads the exchange file and modifies the simulation if needed (A
specific ISIS rule should be created for that matter)
8 ISIS runs next time step and results are produced
9. Back to step 3.
Jean
