20130912 Report : discussion Thursday 20130912 10:30-11:15
Proposal
Need more clarification and reference
Pages . Issues > Feedback < possible answer/solution/todo
Surprise > not academic
Our present day > Should be changed
> More description a bout CCGA including strength and weakness
lower computational cost > explain why? + \cite[who] to support
> add "Scope" section
Next > not linked to previous < another problem
Genetic Algorithms > change to lower case
modification of credit assignment > remove, not necessary to add CCGA-2
constraints > equality \phi(x) and inequality \psi(x) constraints
tough optimization > change "tought" to more academic tone
Big-O notation > remove
> Detain what is n in O(n log n) or O(n^2)
NP-hard problems > never mentionned before, need more detail
Reproduce the following generation > the next population
> add \cite(melanie1999)
> listing of programming procedure is not continued < add an introductory paragraph
Tournament selection > need ref[] or more description < ok + str('because it requires a single pass')
Genetic operator > need figures to describe the process, if copy from books, cite the source.
2.3 CCGA > need description of each procedure,
CCGA-1 > remove '-1', because CCGA-2 is not much in use
stiffness factor > stiffness factors
> add a section about vibration in truss structure,
> explain \phi, [k], [m],
> include figure of truss structure
[m]_i > wrong > local vs. global coordinates
> + figure < ok refer to previous page section about vibration in truss
> fitness = -f
> fitness fucntion is different from objective function
convergence is not known earlier > convergence can be defined earlier
efficiency .... > not clear
fitness criteria > max criteria, in this case we are evaluting minimum criteria, minumum [R]
algo 3.1 line 5 > show relation btw objective function and fitness
need to be redefined (last line) > ?? < same as as page 8 from source, reference to CCGA-2
Algo 3.2 line 2 > initialize an arbitrary \beta_i> is this one or many < a set of predicted damage index [\beta_i]
each case of damage > each case of damage detection or > each case of damage factor
Algo 3.3 l-1 l.+ random noise > Why? + reference , send reference
Algo 3.3 l-8 Eval fitness value>?
Need more detail of parameter settings before the table.
TSME Icome Conference
- approval received 12 Sept
- ask for funding from Dr. Thongchai
- prepare presentation early
Defense
- May be after 20th October
Todo
- Correct proposal
- Send reference: detection with noises
- Meet Dr. Thongchai
- Matlab: clean, make object oriented, add 3D vibration class
- Schedule meating next week 19 Sept 2013, same time is ok