GTOC1: Questions and Answer page

Seventeen international teams started working on a trajectory optimisation problem released by the ACT on the 07/11/2005. Results were sent back by the 5th of December and ranked by the ACT according to the objective function achieved and the constraint violation.

Back to "GTOC1: home page"

Questions and answers: here we have published all the emails received by different teams and the answer we gave them together with some ficticious questions and answers that clarify some points of the competition.

Q: How will the European Space Agency, and in particular your team (the ACT), evaluate the various trajectories received?

A: From the data files received a forward integration will be performed for each of the trajectory phases and the objective function reached will be evaluated again together with a check on the trajectory feasibility.

Q: I 'm not very happy with the last limitation (on the time of flight) which have been decided by ACT team: such decision at this time if it has been discussed with some other teams in competition, must be approved by all the teams if we want a fair competition.
For example my team has already calculated long trajectories to improve as much as possible the cost function and all the work done has been done in this direction! This new limitation induces for our team a lost of time compared to the other teams that have not tried such trajectories. I suggest to ask all the teams what is their actual maximal arrival date and set the maximum arrival date to this value (and not limit the fly duration)

For our team the actual arrival date is in year 2064.I will also suggest an other precision to be put in problem description: Can we define the value of g0 the normalise gravity to be sure that everybody will have the same exhaust velocity for the thruster and so the same mass flow rate. I take g0=9.80665 m/s^2

A: we understand your disappointment, but as we have explained the limit on the time of flight was set for practical purposes and will help us in providing a fair evaluation of the obtained results and in dealing with the data received. We are sorry that your team has already worked on trajectories that are now infeasible and we therefore have decided to provide (to all teams) an extension of four days to send the results back so that the teams will have three weeks from the date of the last modification to the problem. The new deadline is therefore set to be the 4th of December.

Your suggestion on putting a new limit to the arrival date equal to the the maximum arrival date found so far by any of the team is certainly interesting but we decided not to change the rules once again and we think that giving three weeks time from the last major problem modification is a fair compromise.

We have also taken into account your suggestion on the value of g0 and included it as a suggested value in the problem description.

Q: On page 3 of 'The Problem' the Keplerian elements of the asteroid are given for MJD 53600, which is a strange number. Is this correct? (53600 in MJD2000 is the year 2147..).

Also, can you confirm that the thrust is 0.04 N? (and not 0.4 N?)

A: thank you very much for your contribution. First of all I think we need to keep in mind that the optimal solution is not ment to be representative of something that could be flown. Rather, the goal is to test the available softwers for their ability to search for an optimal solution in a complicated (and large) domain. The ultimate goal is to have a workshop here in Estec where we can take the results as starting point for interesting discussion.

This said, you definitely made a point and it led to some talks here within the ACT. To avoid such long solutions which - if proposed - would not be of much interest, we agreed to set a limit on the time of flight of 30 years. The file size should be of about 11000 lines which we believe is manageable. We are also setting up an ftp account where all teams will be able to upload their solutions so to avoid e-mail problems.

We will make a version 3 of the problem description document and add some lines in the "news" section of the competition website (ww.esa.int/act/mission_analysis/globaloptimisationcompetition.htm) make sure to check it from time to time. Hopefully this will be the last change and update.

Q: I am surprised by your answer (refers to the 2100 limit on the arrival date)

If there is no cost to using time, then it is very often found that an optimal trajectory will consume all of the time available. Since launch may be as early as 2010 that could yield a flight time of 90 years-something which is not practical. (You also want the flight data sampled every day for the file to be submitted to you; a flight time of 90 years will thus yield almost 33,000 lines of output.)

I think that the problem is already challenging with such a large launch window. Now with the possibility that only the great-grandchildren of the mission designers would observe the result I think it is more fantastical than representative of something that could be flown.

A: thank you very much for your contribution. First of all I think we need to keep in mind that the optimal solution is not ment to be representative of something that could be flown. Rather, the goal is to test the available softwers for their ability to search for an optimal solution in a complicated (and large) domain. The ultimate goal is to have a workshop here in Estec where we can take the results as starting point for interesting discussion.

This said, you definitely made a point and it led to some talks here within the ACT. To avoid such long solutions which - if proposed - would not be of much interest, we agreed to set a limit on the time of flight of 30 years. The file size should be of about 11000 lines which we believe is manageable. We are also setting up an ftp account where all teams will be able to upload their solutions so to avoid e-mail problems.

We will make a version 3 of the problem description document and add some lines in the "news" section of the competition website (ww.esa.int/act/mission_analysis/globaloptimisationcompetition.htm) make sure to check it from time to time. Hopefully this will be the last change and update.

Q: Thanks for sending this very interesting problem. If I may, I would
like to ask for a clarification about the permitted flight time and
ephemeris: Is there an upper limit on the flight time? Also, what
ephemeris should we use if our trajectory extends beyond the end of
the DE405 ephemeris (ca. Feb 2201). Another question my team has is
whether there is a lower limit on the permitted heliocentric distance
of the spacecraft.

A: thanks for the questions you asked, this gives me an occasion to clarify some important points that we have overlooked.

1 - There is an upper limit on the flight time, no trajectories should extend after 2100.

2 - There is a minimum heliocentric distance allowed of 0.2 Astronomical Units

I am sorry these information were not included in the first problem release, please find attached the revised version of the problem description, also available on the ACT web site www.esa.int/act/mission_analysis/globaloptimisationcompetition