laucalle / angrybirdslevelgenerationthesis Goto Github PK

Decide on a title along the lines of "Free form evolution for angry bird level generation"

LaTeX Warning: Reference `ga:cross2' on page 16 undefined on input line 79.

[16 <./exp2_explication.png>]) (./Conclusion.tex

LaTeX Warning: Reference `ch:context' on page 17 undefined on input line 4.

And some others. Probably left over from original thesis.

Generate .bibs for all papers and books used during the thesis, and add them to a .bib file.

Drawing from conclusions, write the abstract relating it to the title.

And adjust length if needed.

I think experiments 5 and 6 should be fleshed out, but there's still some way to go, mainly devoting more time to any of them.
Also, it would be interesting to try something that is left over from previous versions, try fully built structures (like arcs) as either

• Initial structures
• As an operator that will add them randomly.
  Structures in the latest versions look fine, but they still should look better to be a hosting place for pigs.

Write the introductory chapter, with motivation and general narrative of the thesis

Including bibliography. Same as before, we will have to reduce it significantly. The main part that can be eliminated is the one that refers to EAs.

The state of the art has to be revised, after papers published in CIG and whatever participated in the competition. @Laucalle with help from @amorag, who attended the conference.

It's the best if the intro from your Degree thesis is copied (or at least the fundamental part) to the Introduction Section. Then we can cut it down to desired length.

The methodology chapter must describe, at different levels, the steps you have walked to solve the problem; what you have done at the programming level, and what you have done at the algorithmic and experimental level. Any software engineering/scientific technique you have used should go here.

There are these errors

Warning--to sort, need author or key in ieee-cig
Warning--to sort, need author or key in sciencebirds
Warning--to sort, need author or key in sciencebirds-adapt
Warning--to sort, need author or key in ab-level
Warning--empty booktitle in blum1970stability
Warning--empty booktitle in estevez2017statistical
Warning--can't use both volume and number fields in estevez2017statistical
Warning--empty booktitle in fortin2012deap
Warning--can't use both volume and number fields in fortin2012deap
Warning--empty booktitle in perone2009pyevolve
Warning--can't use both volume and number fields in perone2009pyevolve

And address them.

The state of the art describes what is the current best solution to the level generation problem, and how it has arrived to that state. The purpose of a thesis is to advance the state of the art.

The experiments and results should follow a narrative where, at every step, you formulate a hypothesis, create an experiment to test it, and to explain these results you generate new hypotheses and new experiments, until you find a satisfactory result.

The conclusion should tie with the objecives (from #1) explaining to what extent the objectives have been achieved. If they have not (this is not a PhD), what kind of things should be done to solve them, that is the future work area.
From what has worked and what has not, you should try and draw conclusions. What has worked? Has it worked for the same reason you though it should work? Maybe additional experiments could shed some light on that?

It's IEEE format. Start by copying the latest version (submitted to GECCO).