首页
网站开发
桌面应用
管理软件
微信开发
App开发
嵌入式软件
工具软件
数据采集与分析
其他
首页
>
> 详细
代写data程序、代做C/C++, Java编程设计
项目预算:
开发周期:
发布时间:
要求地区:
Implementation exercises for the course
Heuristic Optimization
Universite Libre de Bruxelles — 2024 ´
Implementation exercise sheet 2
Design and implement two stochastic local search algorithms for the linear ordering problem (LOP), building on
top of the perturbative local search methods from the first implementation exercise. Apply your algorithms to the
same instances as in the first implementation exercise.
The SLS algorithms can be based on either simple, hybrid, or population-based methods chosen among the ones
described in the lectures. The two SLS methods chosen must belong to two different classes. For example,
the first one could be a tabu search (a simple SLS method) and the second one could be an ACO algorithm (a
population-based SLS method). To get inspiration for ways how to generate solutions, operators etc. you may
consider algorithms that have been proposed in the literature. In the articles provided in the “reference” folder in
this channel you will find a comprehensive review of heuristic algorithms for the LOP.
The two algorithms implemented should be evaluated on all instances of size 150 and be compared using statistical
testing. In addition, on few instances the comparison should also be done by measuring run-time distributions. The
implementation should make use of the iterative improvement algorithms from the first implementation exercise.
Please note that the experimental comparison should be among implementations that are coded in the same programming language, the algorithms should be compiled using the same compiler with same compiler flags, and the
experiments should be run on the same computer under same conditions (e.g. no other compute or memory intensive programs are running on the same machine). In other words, the observed differences should be attributable
to differences in the algorithms and not to differences in the experimental conditions.
The deadline for the implementation exercise is: May 5, 2024 (23:59)
Exercise 2.1
1. Run each algorithm once on each instance of size 150. Instances are available on TEAMS and are the same
as used in the first implementation exercise. As termination criterion, for each instance use the average
computation time it takes to run a full VND (implemented in the previous exercise) on the same instance
size and then multiply this time by 100 (to allow for long enough runs of the SLS algorithms).
2. Compute for each of the two SLS algorithms and each instance the mean relative percentage deviation from
the best known solutions.
3. Produce correlation plots of the relative percentage deviation for the two SLS algorithms (see lectures).
4. Determine, using statistical tests (in this case, the Wilcoxon test), whether there is a statistically significant
difference between the relative percentage deviations reached by the two algorithms. Note: For applying the
statistical test, the R statistics software can be used. The software can be download from http://www.
r-project.org/.
5. [Optional] Measure, for each of the two implemented SLS algorithms on three instances qualified runtime distributions to reach sufficiently high quality solutions (e.g. the best-known solutions available or
some solution value close to the best-known one such as 0.1%, 0.25%, or 1% worse than the best-known
solution). Measure the run-time distributions across 25 repetitions using a cut-off time of 10 times the
termination criterion above. As the instances take the first three instances of size 150.
6. Produce a written report on the implementation exercise:
• Include a clear description of the implemented SLS algorithms, mentioning how they work, what are
their algorithm components, and what were the computational results obtained by each of them. Justify
also the choice of the parameter settings that are used and the choice of the method for generating (i)
initial solutions and (ii) the iterative improvement algorithm.
• Present the results as in the previous implementation exercise using tables and statistical tests.
• Present the performance correlation plots and [optionally] present graphically the results of the analysis
of the run-time distributions.
INFO-H413 — Heuristic Optimization 2 Implementation exercise 2
• Interpret appropriately the results (statistical tests, run-time distributions, correlation plots) and make
conclusions on the relative performance of the algorithms across all the benchmark instances studied.
Recommended starting point for literature search:
The article below is the best starting point for exploring the available literature for examples of how
the various existing SLS methods can be adapted to the LOP. We suggest to continue from this article
the search for specific examples of SLS algorithms that have been applied to the LOP. A pdf of this
article can be found in the Files tab of this channel on TEAMS.
• Rafael Mart´ı, Gerhard Reinelt, and Abraham Duarte. A benchmark library and a comparison
of heuristic methods for the linear ordering problem. Computational Optimization and Applications. 51(3):1297-1317 (2012).
Additional information:
• In order to pass the exam, you have to successfully complete this second implementation exercise.
• You need to do the implementation exercise by yourself — cooperation is forbidden.
• You have to submit the following items in a zip folder with your name via TEAMS:
(i) a report in pdf format with scientific article structure (see examples provided on TEAMS)
that concisely explains the implementation, reports the above mentioned tasks (averages,
standard deviations, run-time distributions, correlation plots and results of statistical tests),
and interprets concisely the observed results;
(Note: some of the criteria to be evaluated in the report are: the use of a scientific article structure including
abstract, intro, problem description, material and methods, results, conclusion; the use of tables and plots to
present the obtained results; and the use of statistical tests to draw conclusion about the algorithms performance.)
(ii) the source code of the implementation together with a README file explaining how to compile and/or execute the algorithms from a command line in Linux/MacOS;
(Note: some of the criteria to be evaluated in code are: compilation/execution without errors, the use of structures,
code efficiency, indentation and comments.)
(iii) a simple .txt file with the raw data that was used for statistical testing.
It is important to stress that your code should compile/run on Linux/MacOS without errors and
produce the requested outputs when executed on the provided instances; otherwise the implementation exercise will be considered insufficient.
• As programming language, you may use preferably C, C++, or Java. While using Python for
this exercise is also possible, we recommend against it because it is much more slower than the
alternatives. The sample routines are available only in C++, but you are free to adapt them at
your convenience. Please make sure that your code is properly commented and indented, and
that the README file mentions the exact commands for its compilation and execution.
• Happy coding :-D
软件开发、广告设计客服
QQ:99515681
邮箱:99515681@qq.com
工作时间:8:00-23:00
微信:codinghelp
热点项目
更多
data程序代写、代做c/c++编程语...
2024-05-17
data程序代写、代做python编程...
2024-05-17
program代做、c/c++,python程...
2024-05-17
代写math 3333 3.0 - winter 2...
2024-05-17
代做seng6110 programming ass...
2024-05-17
代写seng6110 object oriented...
2024-05-17
代写comp828: statistical pro...
2024-05-17
代做culture and society调试数...
2024-05-17
代做comp 4911 winter 2024 as...
2024-05-17
代做lh physical iiib / 03 33...
2024-05-17
代做3032ict big data analyti...
2024-05-17
代写comp4702 report代写留学生...
2024-05-17
代写fin2020 hw6代写c/c++编程
2024-05-17
热点标签
fit2004
fit3152
mec208
econ20120
cpt304
econ2101
econ0051
engi4547
econ1048
eengm2510
fit1008
7033mkt
ec2066
cct380h5f
man00019m
mech265001
fin2020
fit9137
n1542
csc4140
math6119
comp1710
fina864
csys5020
busi4412
math5007
2702ict
dts204tc
comp2003j
cosc2673
ecmt2150
bff3121–
comu7000
stat6118
comp814
acc202
ematm0067
bit233
ecs776p
600543
bpln0025
comp3400
econ7030
159.342 ‐ operating
mang6134
math1005/math6005
geog5404m
comp1710/6780
infs 2042
inf6028
bman30702
math0002
msci242l
mgt11001
com00177m
bman71282
fit2001
cpt210
159.341
econ7310
comp3221
comp10002
cpt206
ecmt1010
finm081
econ2005
cpt202
fit3094
socs0030
data7201
data2x01
mn-3507
mat246h1
ib2d90
ib3j80
acc207
comp90007
compx518-24a
fit1050
info1111
acct2201
buad801
compsci369
cse 332s
info1110
math1033
scie1000
eeee2057
math4063
cmt219
econ5074
eng5009
csse2310/csse7231
ec333
econ0001
cpt204
elec4630
ma117
dts104tc
comp2017
640481
csit128
eco000109m
finc5090
ggr202h5f
nbs8295
4ssmn902
chc6171
dsa1002
ebu6304
comp1021
csci-ua.202
com6511
ma416
mec206
iom209
bism7202
idepg001
cpt106
comp1212
ecom209
math1062
mn-3526
fnce3000
fmhu5002
psyc10003
fina2222
be631-6-sp/1
finc2011
37989
5aaob204
citx1401
econ0028
bsan3204
comp9123
cmt218
itp122
qbus6820
ecmt1020
bus0117
soft3202/comp9202
basc0057
mecm30013
aem4060
acb1120
comp2123
econ2151
ecmt6006
inmr77
com 5140
ocmp5328
comp1039
had7002h
cmt309
asb-3715
elec373
cpt204-2324
be631-6-sp
econ3016
mast10007
buss6002
comp4403
comp30023
finm1416
csc-30002
6qqmn971
fin668
mnfg309
inft2031
cits1402
comp2011
eecs 3221
ebu4201
ct60a9600
com336
8pro102
econ7300
comp3425
comp8410
comp222
finm8007
comp2006
comp26020
comp1721
eeen3007j
cis432
csci251
comp5125m
com398sust
finm7405
econ7021
fin600
infs4205/7205
mktg2510-
32022
mth6158
comp328
finn41615
2024
mec302
联系我们
- QQ: 9951568
© 2021
www.rj363.com
软件定制开发网!