Beskjeder

All reports for oblig 1 and 2, which have been e-mailed to me, have been accepted. You may check oblig acceptance here Please let me know if something should be corrected. On the Tuesday Nov 25, I will in addition to summarizing the course and answering questions, also return reports and say a few words about oblig 2.

There will take place a doctoral defence relevant for this cource on Monday Nov 10, 13:15 (don't be late!) as announced here. This is not a compulsory part of the cource but one of the papers in the thesis is included in the syllabus and covered in the lecture Nov 4.

As a part of the lecture October 28, I will say a few words about the solution of Oblig 1. I will also return accepted reports to those who are there. October 28 is also the submission deadline for Oblig 2. Jim

Oblig2: The exercise text has been updated to specify that a round-trip is wanted. If you already have implemented the system with the first and the last city in the route not being the same, this solution is also fine and you do not have to make changes.

Thursday October 16, 09:15-10:00, we will have a guest lecture of Professor Tetsuya Higuchi from Japan. He is regarded as one of originators of evolvable hardware and has for the last ten years together with his group had a huge success in implementing evolvable hardware for real-world applications.

This lecture would be a part of the cource and I hope as many of you as possible are able to attend.

Location: Alfa/Omega room at 4th floor in the Informatics building (up the stairs outside room 3A and turn left)

Oblig 2: Mid-term exercise #2 is now ready and can be dowloaded here . It must be accepted to be allowed to take exam. The report submission deadline is October 28. Contact jimtoer@ifi.uio.no for questions/problems. This is the first version of this exercise so there could be unclear issues.

Oblig 1: Noen har kommentert at n?r ex1 kj?res flere ganger, s? kommer det ut forskjellig resultat hver gang. Det skyldes at initialverdier i populasjonen ved oppstart av programmet blir nye hvis ikke seed ("fr?") er oppgitt. Oppgaven sier ikke noe om dette, s? det blir valgtfritt om en ?nsker ? angi et fr? (ex1 seed <heltall>). Hvis en gj?r det, blir inititialverdiene like og det kan v?re bedre for ? sammeligne forskjellige parametre. (Eng: Every time the program ex1 is executed, a new ¡°seed¡± is generated resulting in new initial genomes in the population. However, it is possible to specify a seed making the initial genomes equal between experiments: ex1 seed <integer>. This approach could be preferable (and you may use it if you like) for in a better way comparing different parameter settings.)

Kursets obligatoriske ?ving nr. 1 er n? klar. Den m? v?re best?tt for f? kunne avlegge eksamen i kurset. Det gis kun godkjent/ikke godkjent p? oppgaven. Leveringsfrist for rapport er 7. okt. Oppgaven lastes ned her . Ta kontakt med jimtoer@ifi.uio.no ved eventuelle problemer eller sp?rsm?l. Compulsory exercise #1 must be approved for being allowed to take exam. Deadline Oct. 7.