--- For your convenience, this form can be processed by EasyChair
--- automatically. You can fill out this form offline and then
--- upload it to EasyChair. Several review forms can be uploaded
--- simultaneously. You can modify your reviews as many times as
--- you want.
--- When filling out the review form please mind
--- the following rules:
--- (1) Lines starting with --- are comments. EasyChair will
--- ignore them. Do not start lines in your review with ---
--- as they will be ignored. You can add comments to the
--- review form or remove them
--- (2) Lines starting with *** are used by EasyChair. Do not
--- remove or modify these lines.
--------------------------------------------------------------
*** REVIEW FORM ID: 81540::51595
*** SUBMISSION NUMBER: 19
*** TITLE: A Discrete Binary Version of the Electromagnetism-Like Heuristic for Solving the Combinatorial Optimization Problems
*** AUTHORS: Nikbakhsh Javadian, Mohsen Gol Alikhani and Reza Tavakkoli-Moghaddam
*** PC MEMBER: Todd Wareham
--------------------------------------------------------------
*** REVIEW:
--- Please provide a detailed review, including justification for
--- your scores. This review will be sent to the authors unless
--- the PC chairs decide not to do so. This field is required.
This paper proposes a discretized Electromagnetism-Like (EM) heuristic for
solving combinatorial optimization problems. This heuristic modifies an
existing EM heuristic by applying a variant of the discretization method
previously applied to particle-swarm optimization systems. This heuristic
is applied to the Traveling Salesman Problem (TSP), and in testing appears
to perform better (in terms of closeness of solutions to optimal) than
previous EM algorithms for TSP.
This paper is well-organized and well-written; there are some
minor grammatical errors (some of which are noted below), but these do not
detract from understanding the paper. The heuristic is well-described, and
the previous work on which it is based (and the precise nature of
this relationship) is nicely and honestly described as well. The testing,
though extensive and well-described, is not quite systematic enough to
convince me of the magnitude of the improvement given by the new heuristic;
however, it certainly does appear to perform better than previous EM
algorithms in the cases examined.
To summarize, this is a well-written paper describing a modification adopted
from an existing heuristic algorithm of another heuristic algorithm that
appears to result in improved performance, though the given testing does not
firmly establish the magnitude of this improvement. Hence, I recommend
neither acceptance nor rejection wrt COCOA.
General comments:
- Multiple citations should be in the same square brackets, i.e.,
replace "\cite{X} and \cite{Y}" by "\cite{X, Y}".
- "... the continuous space ..." => "... continuous space ..."
- What was the running times of the various EM algorithms described in
this paper? How do these running times (and the qualities of
obtained solutions) compare with those for non-EM heuristics for TSP?
This should be mentioned in the Testing section (Section 5).
Specific comments:
- Title: ... Solving the Combinatorial ..." => "... Solving
Combinatorial ..."
- Abstract, last line: "... the previous ..." => "... algorithms in
previous ..."
- Page 1, paragraph 1, line 5: "... can be converged ..." =>
"... converges ..."
- page 2, paragraph 2, line 5: "... Kennedy and Eberhardt ..." =>
"... Eberhardt and Kennedy ..."
- page 2, paragraph 3, line 2: "... used in ..." => "... used to
solve ..."
- page 2, paragraph 3, last 2 lines: "... the TSP in large sizes." =>
"... large instances of TSP."
- page 3, paragraph 1, last line: "... EM and ... step." => "... EM:"
- page 4, paragraph 2, line 1: "In the ..." => "The ..."
- page 4, paragraph 2, line 2: "... the premature ..." =>
"... premature ..."
- page 4, end of paragraph 2: Is global convergence with probability one
after an infinite / non-polynomial number of iterations? This should
be clarified.
- page 7, paragraph 1, last sentence: This statement is a bit strong --
Table 6 shows that for particular datasets, the instances cannot be
solved exactly, but are solved fairly closely (especially by [5]).
Perhaps it should be said that solution quality appears to drop as
number of cities increases. This should be verified by (1) testing
on instances with higher number of cities (preferably those instances
used in the results described in Table 7) and (2) re-running the
algorithms from [4] and [5] with higher numbers of iterations
(comparable to those used in Table 7), especially for the larger
inputs. The proposed EM should also be run on the test cases used in
Table 6. Doing all this would allow a fairer comparison of all EM
algorithms (how does the proposed discretized EM work on small
instances? / can the older EM converge to better solutions if a
comparable (for the proposed EM) number of iterations is allowed?)
In addition, to verify that observed behavior holds for inputs of
a particular size in general, more than one input of each size should
be run and the average of the best solutions obtained for all
instances of a given size should be graphed.
- page 7, paragraph 2, line 7: Shouldn't 500 be 400 here (based on what I
see in Table 7)?
- page 8, paragraph 1, line 1: "... lager ..." => "... larger ..."
- page 8: Figures 1 and 2 could collapse into one figure with
3 plot-lines. Even if this is not done, effort should be made to
distinguish these lines (one dashed / one solid?), and "WU" should be
replaced by "Wu" in the captions.
--------------------------------------------------------------
*** REMARKS FOR THE PROGRAMME COMMITTEE:
--- If you wish to add any remarks for PC members, please write
--- them below. These remarks will only be used during the PC
--- meeting. They will not be sent to the authors. This field is
--- optional.
--------------------------------------------------------------
--- If the review was written by (or with the help from) a
--- reviewer different from the PC member in charge, add
--- information about the reviewer in the form below. Do not
--- modify the lines starting with ***
*** REVIEWER'S FIRST NAME: (write in the next line)
*** REVIEWER'S LAST NAME: (write in the next line)
*** REVIEWER'S EMAIL ADDRESS: (write in the next line)
--------------------------------------------------------------
--- In the evaluations below, uncomment the line with your
--- evaluation or confidence. You can also remove the
--- irrelevant lines
*** OVERALL EVALUATION:
--- 3 (strong accept)
--- 2 (accept)
--- 1 (weak accept)
0 (borderline paper)
--- -1 (weak reject)
--- -2 (reject)
--- -3 (strong reject)
*** REVIEWER'S CONFIDENCE:
--- 4 (expert)
--- 3 (high)
2 (medium)
--- 1 (low)
--- 0 (null)
*** END
--------------------------------------------------------------
--------------------------------------------------------------
--- For your convenience, this form can be processed by EasyChair
--- automatically. You can fill out this form offline and then
--- upload it to EasyChair. Several review forms can be uploaded
--- simultaneously. You can modify your reviews as many times as
--- you want.
--- When filling out the review form please mind
--- the following rules:
--- (1) Lines starting with --- are comments. EasyChair will
--- ignore them. Do not start lines in your review with ---
--- as they will be ignored. You can add comments to the
--- review form or remove them
--- (2) Lines starting with *** are used by EasyChair. Do not
--- remove or modify these lines.
--------------------------------------------------------------
*** REVIEW FORM ID: 83931::51595
*** SUBMISSION NUMBER: 30
*** TITLE: Linear Ordering Problem Applications and Extensions
*** AUTHORS: Philippe Mahey, Mourad Baiou and Abilio Lucena
*** PC MEMBER: Todd Wareham
--------------------------------------------------------------
*** REVIEW:
--- Please provide a detailed review, including justification for
--- your scores. This review will be sent to the authors unless
--- the PC chairs decide not to do so. This field is required.
This paper presents a collection of research proposals (two of which have
associated algorithmic results) for the linear ordering problem, which has
applications in economics. The two results are (1) a revised integer linear
programming (ILP) formulation for the general problem (to aid in finding
better exact solutions in general) and (2) an ILP formulation for a
practically useful restricted version of the general problem, i.e.,
fixed-cardinality linear ordering. Both formulations are tested relative to
existing benchmark datasets using the ILP solver CPLEX, and results are
given and interpreted. Additional algorithms for solving both formulations
are also briefly sketched.
This paper is well-written and well-organized, the linear-ordering problem
is defined clearly, and previous algorithms are summarized well. The
proposals for additional algorithms also seem interesting. However, the
paper would be much stronger to me if it more fully described and actually
implemented the proposed algorithms for solving the given ILP formulations.
As such, it seems much more a proposal of work to do rather than a
summary of work done.
To summarize, this is a well-written paper describing preliminary work. As
such, I cannot recommend it for acceptance at COCOA.
Specific comments:
- Abstract, line 2: "It also ..." => "It ..."
- Abstract, line 5: "Additionally, an ..." => "An ..."
- Page 2: What is the computational complexity of the linear ordering
problem? can it be solved exactly in polynomial time? If it NP-hard?
I suspect the latter, but that should be clearly stated somewhere
(with a citation to the literature backing this up).
- Page 4, paragraph 2: What is a percentage duality gap?
- Page 7, paragraph 2: More discussion of results might be appropriate
here. Also, what are the instances xxx and yyy mentioned in
the last line of this paragraph?
--------------------------------------------------------------
*** REMARKS FOR THE PROGRAMME COMMITTEE:
--- If you wish to add any remarks for PC members, please write
--- them below. These remarks will only be used during the PC
--- meeting. They will not be sent to the authors. This field is
--- optional.
--------------------------------------------------------------
--- If the review was written by (or with the help from) a
--- reviewer different from the PC member in charge, add
--- information about the reviewer in the form below. Do not
--- modify the lines starting with ***
*** REVIEWER'S FIRST NAME: (write in the next line)
*** REVIEWER'S LAST NAME: (write in the next line)
*** REVIEWER'S EMAIL ADDRESS: (write in the next line)
--------------------------------------------------------------
--- In the evaluations below, uncomment the line with your
--- evaluation or confidence. You can also remove the
--- irrelevant lines
*** OVERALL EVALUATION:
--- 3 (strong accept)
--- 2 (accept)
--- 1 (weak accept)
--- 0 (borderline paper)
-1 (weak reject)
--- -2 (reject)
--- -3 (strong reject)
*** REVIEWER'S CONFIDENCE:
--- 4 (expert)
--- 3 (high)
2 (medium)
--- 1 (low)
--- 0 (null)
*** END
--------------------------------------------------------------