-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathCE310_Coursework.html
950 lines (739 loc) · 109 KB
/
CE310_Coursework.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
<!DOCTYPE html>
<html>
<head>
<meta charset="UTF-8">
<!-- #comment-guard: (Join the two dashes below to enable auto-refreshing; remember to BREAK THEM APART AGAIN FOR THE FINAL DOCUMENT)
- -> <meta http-equiv="refresh" content="2" />
<!-- #end-comment-guard -->
<title>CE310 Coursework - Spring Term 2023-2024</title>
<script async type="text/x-mathjax-config">MathJax.Hub.Config({tex2jax: {inlineMath: [ ['[$]','[/$]'] ], displayMath: [ ['[$$]','[/$$]'] ], skipTags: ["script","noscript","style","textarea", "code"] }});</script>
<script async type="text/javascript" src="http://cdn.mathjax.org/mathjax/latest/MathJax.js?config=TeX-AMS-MML_HTMLorMML"></script>
<style>
a { color: #ffddaa; }
b { color: #66ff66; }
blockquote { background-color: #ffffe0; font-style: italic; font-family: "Gentium"; padding: 1em; color: #000000; }
body { font-family : "Helvetica", sans; font-size : 13pt; text-align : center; background-color: #222222; color : #eeeeee; }
h1 { margin-top: 2em; margin-bottom: 1em; }
h2 { margin-top: 3em; margin-bottom: 1em; }
h3 { margin-top: 3em; margin-bottom: 1em; font-style: italic; }
h4 { margin-top: 2em; margin-bottom: 1em; font-style: italic; }
center { display: block; text-align: center; margin-top: 1em; margin-bottom; 1em; }
flex { display: flex; justify-content: center; margin-top: 1em; margin-bottom; 1em; }
html { scroll-behavior: smooth; }
img { display: inline-block; }
li { margin-bottom: 0.75em; margin-top: 0.5em; }
q { font-style: italic; font-family: "Gentium"; }
table { border-collapse: collapse; }
th { border: 1px solid white; border-top: 3px solid; border-bottom: 3px solid; padding: 0.5em; background-color: #666666; font-weight: bold; }
tr:nth-child(odd) { background-color: #303030; }
tr:nth-child(even){ background-color: #404040; }
td { border: 1px solid white; padding: 0.5em; }
tt, pre, code { background-color: #282833; color: #f4f4ff; font-size: 0.9em; font-family: monospace; padding: 0.25em; border-radius: 0.5em; line-height: 1.5em; overflow: auto; }
tt, code { padding-left: 0.5em; padding-right: 0.5em; }
blockquote b { color: #110011; }
blockquote tt { background-color: rgba(20%, 20%, 30%, 0.25); color: #000000; font-style: normal; }
summary { font-size: 1.2em; font-weight: bold; margin-top: 1.5em; margin-bottom: 0.5em; }
centralcontainer { display : inline-block; width : 60em; margin-bottom: 5em; text-align : left; }
toc { background-color: #444444; border: 1px solid black; border-radius: 1em; padding: 1em; padding-left: 2em; display: block; }
hint { margin-left: 3em; }
div.markscontainer { display: block; text-align: right; font-size: 0.8em; }
div.panel { display: inline-block; border: 2px solid black; border-radius: 0.5em; padding-left: 1em; padding-right: 1em; background-color: #444444; }
/* colours */
rc { color: #ff4444; } rb { color: #ff4444; font-weight: bold; }
oc { color: #ffaa66; } ob { color: #ffaa66; font-weight: bold; }
gc { color: #66ff66; } gb { color: #66ff66; font-weight: bold; }
bc { color: #66aaff; } bb { color: #66aaff; font-weight: bold; }
pc { color: #aa66ff; } pb { color: #aa66ff; font-weight: bold; }
</style>
</head>
<body>
<centralcontainer>
<!-- ============================ -->
<!-- Start of assessment document -->
<!-- ============================ -->
<!-- Page "Header" -->
<flex>
<div style="width:50%; text-align:left;">CE310: ‘Evolutionary Computation and Genetic Programming’<p style="margin-top: 0.5em;">Coursework (Parts 1 and 2)</p></div>
<div style="width:20%;"></div>
<div style="width:30%; text-align:right;">University of Essex, School of CSEE<p style="margin-top: 0.5em;">Tasos Papastylianou,<br>Spring Term 2023-2024</p></div>
</flex>
<h1><a href="#" style="text-decoration: none;">CE310 Coursework</a></h1>
<hr>
<toc>
<ul>
<li>
<a href="#Overview"> Overview </a>
</li>
<li>
<a href="#Part1"> Part 1 - Implementing basic Evolutionary Algorithms </a>
<ul>
<li> <a href="#Part1_AssignmentObjective"> Assignment Objective </a> </li>
<li> <a href="#Part1_Task1"> Task 1 - Implement a steady-state binary GA library </a> <span style="margin-left: 2em;">[12.5 marks]</span> </li>
<li> <a href="#Part1_Task2"> Task 2 - Solve combinatorial optimisation problems </a> <span style="margin-left: 2em;">[37.5 marks]</span> </li>
<ul>
<li> <a href="#Part1_Task2_Problem1"> Problem 1: ASCII Art </a> </li>
<li> <a href="#Part1_Task2_Problem2"> Problem 2: Hyper-parameter Analysis </a> </li>
<li> <a href="#Part1_Task2_Problem3"> Problem 3: Fitness functions </a> </li>
<li> <a href="#Part1_Task2_Problem4"> Problem 4: The Santa Claus Riddle </a> </li>
</ul>
</li>
<li> <a href="#Part1_Notes_and_guidance" style="margin-top: 1em;"> Notes and guidance on Part 1 </a> </li>
</ul>
</li>
<li>
<a href="#Part2">Part 2 - Genetic Programming mini-project</a> <span style="margin-left: 2em;">[50 marks]</span>
<ul>
<li> <a href="#Part2_AssignmentObjective"> Assignment Objective </a> </li>
<li> <a href="#Part2_TaskDescription"> Task Description </a> </li>
<li> <a href="#Part2_MarkingCriteria"> Marking Scheme for Part 2 </a> </li>
<li> <a href="#Part2_Notes_and_guidance"> Notes and guidance on Part 2 </a> </li>
</ul>
</li>
<li>
<a href="#ProblemsEtc">Problems with the assignment, Late Submissions, Extenuating Circumstances, Academic Offenses, etc.</a> </li>
</li>
</ul>
</toc>
<hr>
<!-- -------- -->
<!-- Overview -->
<!-- -------- -->
<h2 id="Overview">Overview.</h2>
<p> The coursework consists of two parts, each carrying 50 marks (i.e. a
total of 100 marks for both parts).
<br>
This coursework is worth 40% of the module overall (the other 60% being
exams).
<ul>
<li> Part 1 involves coding your own Genetic Algorithm framework in
Python, and using it to solve a small number of optimisation
problems, documenting and discussing your work along the way.
</li>
<li> Part 2 is a mini-project / research task, which asks you to
perform computational Genetic Programming experiments in Python, and
produce a report that discusses the results obtained.
</li>
</ul>
</p>
<p> This document was <a href="https://moodle.essex.ac.uk/course/view.php?id=3666§ion=3#assignment">released on Moodle</a>,
on 2024-01-29 (Week 18).
</p>
<p> The deadline for submitting the coursework on FASER" is 2024-03-08
(Week 23).<br><rb>Note: deadline is at 2pm, not midnight!</rb>
</p>
<p> <b>Submission:</b> please submit all your work as a single zip file on
<a href="https://faser.essex.ac.uk">FASER</a> by the above deadline.
Inside the zipfile, put all Part 1 related material (i.e. code and
documentation / reports) into a folder called <tt>Part1</tt>, and all
Part 2 related material in a folder called <tt>Part2</tt>. Please also
provide a <tt>requirements.txt</tt> file with your submission, so that I
can confirm what python package versions you used in case of problems;
you can obtain this from <tt>pip</tt> via the <tt>freeze</tt> command.
</p>
<p> If you have any questions, feel free to contact me <a
href="mailto:[email protected]">by email</a>, or,
preferably, if you think your question may be of interest/benefit to
your colleagues, please feel free to post it on the <a
href="https://moodle.essex.ac.uk/mod/forum/view.php?id=1203776">Help
Forum on Moodle</a>.
</p>
<!-- ------ -->
<!-- Part 1 -->
<!-- ------ -->
<h2 id="Part1">Part 1 - Implementing basic Evolutionary Algorithms</h2>
<h3 id="Part1_AssignmentObjective">Assignment Objective</h3>
<p> The objective of part 1 is to familiarize yourself with evolutionary
algorithms; specifically, problem solving using <b>genetic
algorithms</b> (GAs).
</p>
<!-- Task 1 -->
<h3 id="Part1_Task1"> Task 1 - Implement a steady-state binary GA library </h3>
<p> Design and implement the core functions needed for <b>a steady-state
binary GA</b> and use the implemented GA to solve combinatorial
optimisation problems (see below for more details). Software
architecture, design and development decisions are entirely up to you
(as long as it's in <rb>Python</rb>, and in the form of a
<rb>package</rb> or <rb>module</rb> that can be subsequently imported
and used in your code). Select and implement <b>crossover</b>,
<b>mutation</b>, and <b>selection</b> operators appropriate for the
problems to be solved. Select appropriate standard values for the
hyper-parameters (population size, generations, crossover rate, etc.).
</p>
<div class="markscontainer"><div class="panel"><p>Marking criteria: code and code documentation [12.5 marks]</p></div></div> <br><br>
<blockquote style="background-color: #00bbff;"> Note that there is no need
to install any additional or specialised libraries for this task.
Software libraries included in the standard Python installation should
be more than sufficient. Having said that, you might want to use things
like <tt>Matplotlib</tt> to visualise results as graphs, or
<tt>NumPy</tt> if you plan to implement more efficient array operations,
etc.
</blockquote>
<blockquote style="background-color: #ffaa44;">
<b>> Not sure how to start? Little or no experience with Python?</b>
<p> Note that in the lecture we will be going through examples
discussing possible implementation details for this kind of framework.
Feel free to refer to code examples from class for inspiration /
reference; (in particular any 'main' functions or equivalent parts of
the code, which implement the main logic of a GA system).
</p>
However, please do not just copy verbatim, since copy-pasting and
then trying to make in-place alterations to code intended for specific
problems is rarely a good way to attack new problems. Instead, study
the logic of the code, and try to create your own similarly.
</blockquote>
<!-- Task 2 -->
<br><br>
<h3 id="Part1_Task2"> Task 2 - Solve combinatorial optimisation problems </h3>
<p> You will now need to write code that imports your module (or package)
from Task 1, and uses it appropriately to solve the optimization
problems below.
</p>
<p> Additionally, you must make observations about GA performance
and hyper-parameters, answer questions about the various problems and
make observations about GA performance and hyperparameter settings for
each problem. <bc>Observations to be made/questions to be answered are
written in blue.</bc>):
</p>
<!-- Problem 1 -->
<br><br>
<h4 id="Part1_Task2_Problem1"> Problem 1: ASCII Art </h4>
<ul>
<li>
<p> Test your code above by using a binary array or bitstring
representation (<b>chromosome</b>) and evolve the representation to
match a given ASCII art image, such as the one in Fig. 1.
</p>
<div style="margin-top: 4em; margin-bottom: 4em;">
</code>
<flex><div style="font-size: 1.6em;">
<span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><br>
<span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><br>
<span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><br>
<span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><br>
<span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #444488;">0</span><span style="color: #aaffaa;">1</span><span style="color: #444488;">0</span><br>
</div></flex>
<flex><div style="font-size: 0.8em;"> Fig. 1. ASCII art 'image' example. The 1s form the letters GA (coloured here for extra visual effect). </div></flex>
</code>
</div>
<p> The specification for this kind of ASCII art is as follows: a text
block where all rows have exactly the same number of columns (i.e.
characters), and where the only valid characters are the symbols
<tt>0</tt> and <tt>1</tt> (except also for the impicit newline
characters <tt>\n</tt> at the end of each column). You may use the
ASCII art textblock from Fig 1., or feel free to create your own.
</p>
<blockquote style="background-color: #aaaa88;"> Hint: Fig 1. is not
actually an <tt><img></tt> element, it's just colored text.
Feel free to copy/paste it directly into your code, e.g. inside a <a
style="color: #222266;"
href="https://stackoverflow.com/q/23361171/4183191">triple-quoted
string</a>.
</blockquote>
<br>
<p> Each line in the ASCII text block represents a row of pixels.
Multiple lines of <tt>0</tt>s and <tt>1</tt>s need to be
concatenated to form a complete binary representation
(<b>genotype</b>) of the 2-D ASCII art matrix (<b>phenotype</b>).
The opposite operation is needed to convert genotypes evolved by the
GA into phenotypes. To evolve solutions, use the following fitness
function:
</p>
<p style="margin-top: 3em; margin-bottom: 3em;">
[$$] \begin{array}{l} \displaystyle f( I ) = \sum_{i \in I} ~g(i; ~ I,A) \\[1em] \text{, where:}~~~ g(i; ~ I, A) = \left \{ \begin{array}{l} 1, ~~\text{if}~~ I(i) = A(i) \\ 0, ~~\text{if}~~ I(i)\neq A(i) \end{array} \right . \end{array} [/$$]
</p>
<p>
(or, in pseudocode):
</p>
<pre><code>
fitness( I ) = sum( [I(i) == A(i)] forall i in I )
</code></pre>
<p> where <tt>I</tt> is an individual to be evaluated, <tt>A</tt> is
the target ASCII-art binary pattern we want to evolve and <tt>i</tt> is the
i<sup>th</sup> gene (bit) in the individual.
</p>
<div class="markscontainer"><div class="panel"> <p> Marking criteria: Code and code documentation [2 marks] </p></div></div> <br><br>
</li>
<li>
<p> Perform different GA runs, i.e., execute your code several times, and
evolve individuals by firstly <b>maximising</b> and secondly
<b>minimising</b> the above fitness function. <bc>What solutions do you expect when maximising and
minimising the fitness function?</bc> (Marking criteria: Observations - 0.5/2.5 marks)
</p>
<div class="markscontainer"><div class="panel"> <p> Marking criteria: Observations [0.5 marks] </p></div></div> <br><br>
</li>
</ul>
<!-- Problem 2 -->
<h4 id="Part1_Task2_Problem2"> Problem 2: Hyper-parameter Analysis </h4>
<ul>
<li>
<p> To get a better picture about the GA performance, experiment with the
GA <b>hyper-parameters</b> such as <b>population size</b>,
<b>generations</b>, <b>crossover</b> <b>rate</b> and <b>mutation</b>
<b>rate.</b> Also explore hyper-parameters (if any) of the
<b>selection</b> <b>operator</b> that you decided to use. Start with
standard values that we discussed in the lectures.
</p>
<div class="markscontainer"><div class="panel"> <p> Marking criteria: Code and code documentation [3 marks] </p> </div></div> <br><br>
</li>
<li>
<p> <bc> Briefly describe and discuss the behaviour of the GA based on
the selected hyper-parameters. Which parameter combination achieves
optimal performance? Why do you think is the parameter combination
successful? Are there any drawbacks when using the identified
parameter combination? </bc>
</p>
<div class="markscontainer"><div class="panel"> <p> Marking criteria: Observations and discussion [7 marks] </p> </div></div> <br><br>
</li>
</ul>
<!-- Problem 3 -->
<h4 id="Part1_Task2_Problem3"> Problem 3: Fitness functions </h4>
<p> Generally speaking, a GA's task is to find the minimum or maximum of a
fitness function, under the assumption that this fitness function is
useful and appropriate for characterising the problem at hand. However,
finding suitable fitness functions that are tractable and can
characterise a problem well can be a challenge. In general, we do not
know the structure (i.e., the shape of the ‘graph’) of the fitness
function; e.g. in principle, we cannot know in advance how many local or
global minima and maxima it has.
</p>
<p> In this task we explore the impact of different representations and
fitness functions (see below) on the performance of your GA. Remember
that optimization involves finding the inputs [$] \mathbf{x} = (x_1,
\dots, x_N) [/$] to a fitness (or more generally objective) function [$]
f( \mathbf{x} ) [/$] that result in the minimum or maximum output of [$]
f ( \mathbf{x} ) [/$]. To solve this task, you need to define an
appropriate representation for [$] x_i [/$] (e.g., binary, real-valued
or mixed) and then evolve this representation (by adapting or creating
binary or real-valued genetic operators as needed) to find the global
minimum or maximum of [$] f ( \mathbf{x} ) [/$].
</p>
<ul>
<li>
<p> Select <b>one</b> single-objective and <b>one</b> constrained
optimisation function from <a href="https://en.wikipedia.org/wiki/Test_functions_for_optimization">https://en.wikipedia.org/wiki/Test_functions_for_optimization</a>
and implement them as your fitness functions. Then run simulations
to find the minimum (or maximum, depending on the problem) of the
function.
</p>
<div class="markscontainer"><div class="panel"> <p> Marking criteria: Code and code documentation [5 marks] </p> </div></div> <br><br>
</li>
<li>
<p> <bc >Briefly describe and discuss the behaviour of the GA. Which
representation did you use? Did the GA find the minimum or maximum?
Did you have to adjust the hyper-parameters? How precise is the
solution? How can you improve the solution? </bc>
</p>
<div class="markscontainer"><div class="panel"> <p> Marking criteria: Observations and discussion [5 marks] </p> </div></div> <br><br>
</li>
</ul>
<!-- Problem 4 -->
<h4 id="Part1_Task2_Problem4"> Problem 4: The Santa Claus Riddle </h4>
<p> Consider the following equation: </p>
<p style="margin-top: 3em; margin-bottom: 3em;">
[$$] \begin{array}{lllllll}
& & S & A & N & T & A \\
- & & C & L & A & U & S \\[0.5em]
\hline
= & & & X & M & A & S \\
\end{array}
[/$$]
</p>
<p> There exists a (non-unique) mapping [$] \mathcal L \mapsto \mathcal D
[/$] from letters in the set [$] \mathcal L = \{ A, C, L, M, N, S, T, U,
X \} [/$] to digits in the set [$] \mathcal D = \{ 0 \dots 9 \} [/$],
such that the above equation holds true, and such that no two letters
are mapped to the same digit*
</p>
<p style="font-size: 0.8em;"> *(in mathematical terms, we would say that [$]
\mathcal L \mapsto \mathcal D [/$] is <a
href="https://en.wikipedia.org/wiki/Injective_function"><i>injective</i></a>).
</p>
<p style="margin-top: 3em;"> E.g. if the mapping [$] \mathcal L \mapsto \mathcal D [/$] was as follows: </p>
<p style="font-size: 0.8em;">
[$$]
\left \{
\begin{array}{ccc}
A & \mapsto & 1 \\
C & \mapsto & 2 \\
L & \mapsto & 3 \\
M & \mapsto & 4 \\
N & \mapsto & 5 \\
S & \mapsto & 6 \\
T & \mapsto & 7 \\
U & \mapsto & 8 \\
X & \mapsto & 9
\end{array}
\right \}
[/$$]
</p>
<p> then this would have resulted in the following equation: </p>
<p>
[$$] \begin{array}{lllllll}
& & 6 & 1 & 5 & 7 & 1 \\
- & & 2 & 3 & 1 & 8 & 6 \\[0.5em]
\hline
= & & & 9 & 4 & 1 & 6 \\
\end{array}
[/$$]
</p>
<p> However, this is equation is clearly false, so this isn't the right mapping. </p>
<ul>
<li>
<p> Solve this problem using a GA approach. You will have to decide on
an appropriate representation for the mapping, and an appropriate
fitness function.
</p>
<div class="markscontainer"><div class="panel"> <p> Marking criteria: Code and code documentation [6 marks] </p> </div></div> <br><br>
</li>
<li>
<p> <bc> Discuss your choice of representation and fitness function. In
your discussion, make sure you cover the following points: </bc>
<ul>
<li> <bc> Did you try to ensure 'uniqueness' in the letter-to-digit
mappings in your representation? If so, how? If not, why? </bc>
</li>
<li> <bc> One possible fitness function is [$$] f( \mathcal L
\mapsto \mathcal D ) = \Biggl | ~ \Bigl ( SANTA-CLAUS \Bigr ) -
XMAS ~ \Biggr | [/$$] Is this a good fitness function? Why / why
not? How could it be improved? If you used a different fitness
function, discuss (and <i>show!</i>) how it compares to this
one, and why you think yours might be a more appropriate fitness
function for this problem. </bc>
</li>
</p>
<div class="markscontainer"><div class="panel"> <p> Marking criteria: Observations and discussion [9 marks] </p> </div></div> <br><br>
</li>
</ul>
<!-- Notes and guidance -->
<h3 id="Part1_Notes_and_guidance" style="margin-top: 1em;"> Notes and guidance on Part 1 </h3>
<blockquote style="background-color: #ffaadd;">
<h3 style="margin-top: 1em;"> Notes and guidance on Part 1 (GA implementation and optimization problems)</h3>
<ul>
<li>
<u><b>Clarification of Marking criteria for Part 1:</b></u><br>
<ul>
<li> <b>Code and code documentation:</b> To get
full marks as listed above, the code must execute without
errors, and reflect source code of high quality, i.e. properly
documented via suitable function/class docstrings, the right
amount of insightful, explanatory comments and sections,
self-documenting names, meaningful code structure(s), and being
a maintainable/readable (to an English reader!) codebase in
general.
</li>
<li>
<b>Discussion-based parts:</b> You must critically answer all
questions posed in the various problems, make observations about
GA performance and influence of hyperparameter settings, and
discuss each problem, providing useful insights as necessary.
</li>
</ul>
</li>
<li style="margin-top: 2em;">
<u><b>What must be submitted to FASER?</b></u><br>
Please submit all your work as a single zip file; the zip file
should split the work into two folders, appropriately labelled
<tt>Part1</tt> and <tt>Part2</tt>. Each folder should contain the
related material (i.e. code and documentation / reports) in a manner
that makes it easy and intuitive for the assessor to discover your
work and run your code. Together with your code, please also provide
a <tt>requirements.txt</tt> file (such as the one that can be
generated via the <tt>pip freeze</tt> command), so that I can
confirm what python package versions you used in case of problems.
Your reports can be produced in any format you prefer and are
comfortable with: .docx, .odt, .pdf, .html, .md/markdown etc are all
fine, as long as I'm able to read them and they are
structured appropriately.
<p> <b style="font-size: 1.1em;">Note:</b> in previous years,
previous module supervisors expected submissions in the form of a
jupyter notebook. Presumably, this was to encourage a “<a
style="color: #222266;"
href="http://www.literateprogramming.com/index.html">literate
programming</a>” approach to your work, to encourage you to think
“out-loud” and document your code as you go along. However, I am
not a big fan of jupyter notebooks, as I feel they encourage bad
programming habits, and so I discourage their use this year.
Nevertheless, <b style="font-size: 1.1em;">do</b> by all means
feel free to use appropriate comment headers / sections in your
code to explain things in a “literate programming” fashion if you
so wish (and indeed, insightful explanations in-code, alongside
clean, self-documenting practices makes for excellent code!). Code
and reports will be marked independently as per the marking
criteria; but, if you do adopt a ‘literate programming’ style in
your code, don't worry about duplicating information in the report
that you have already discussed in the code; as long as both stand
in their own right as independent pieces of work and the extent of
the information provided is appropriate for each particular
context, then any duplication of content between code and report
will not be penalised. On the other hand, naturally, don't rely on
in-code documentation for things that will be expected (according
to the marking criteria) to be found in the report!
</p>
</li>
<li style="margin-top: 2em;">
<u><b>How to start?</b></u><br>
This depends on how familiar you are with algorithmic thinking
and on your coding skills. A good way to start is to make sure
that the task is clear. This includes reading the whole assignment
and making sure you are familiar with the content discussed in
units 1-3.
<p> Check the learning material on Moodle, use resources available
at the university library or use internet search engines to find
relevant material. Of course, asking questions during the lectures
and classes is another good way to get answers and have a good
discussion. I would particularly encourage you to use the <a
style="color: #222266;"
href="https://moodle.essex.ac.uk/mod/moodleoverflow/view.php?id=1203777">MoodleOverflow</a>
facility, to engage your peers in discussion. You can assume that
if you struggle, your fellow students will also struggle. So, I
encourage you to ask questions and engage in constructive
conversations.
</p>
<p> Once it is clear what you need to do start with the planning.
You can start top-down, i.e., plan the development by staring from
the general Evolutionary Algorithm and then implement the
different functions, or bottom-up, i.e., implement basic
functionality and then combine everything. Be systematic and add
functionality step-by-step. Also, be sure to define the way you
want to represent binary arrays/string before implementing the
genetic operators.
</p>
<p> Keep in mind that there are many ways how to address and
complete the tasks. There is no single correct or best way to do
this. Primarily this assignment serves as a learning opportunity;
learning means making errors in a safe environment, and thinking
how to correct them. So above all, start this process <b
style="font-size: 1.2em; color: #aa0000;">early(!)</b>, so that
you have time to try things and make mistakes, and have still have
time to learn from them, comfortably start again and make
progress!
</p>
</li>
</ul><br>
</blockquote>
<br>
<flex><img src="data:image/png;base64,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" alt="[Floral Divider]"></flex>
<!-- ------ -->
<!-- Part 2 -->
<!-- ------ -->
<h2 id="Part2">Part 2 - Genetic Programming mini-project </h2>
<h3 id="Part2_AssignmentObjective">Assignment Objective </h3>
<p> The objective of part 2 is to familiarize yourself with <b>genetic
programming</b> (GP) by running several experiments (symbolic regression
problems, see the teaching material on Moodle or <a
href="https://en.wikipedia.org/wiki/Symbolic_regression">Wikipedia</a>
for more details).
</p>
<p> While part 1 focused on the implementation of the basic algorithms to
gain insights in the underlying mechanisms of evolutionary computation,
part 2 focuses on <b>enhancing</b> your <b>research</b> and <b>analysis skills</b>. More
precisely, like a <b>scientist</b> you will <b>collect</b> and <b>analyse</b> empirical <b>data</b>,
<b>summarise</b>, and <b>interpret</b> your <b>results</b>, and based on the <b>evidence</b>
gathered draw <b>conclusions</b>.
</p>
<h3 id="Part2_TaskDescription"> Task Description </h3>
<p> You are asked to perform a series of GP runs and describe the results
of your runs in a report. In your experiments you will need to use GP in
different configurations (i.e., problems and parameters). Since GP is a
stochastic searcher, you will see that performance varies from run to
run. Therefore, to draw your conclusions, you should ensure you perform
<rb style="font-size: 1.1em;">at least 10 runs</rb> in each
configuration. The aim of the experiments is to get an intuition on how
the population and tournament size impacts on fitness and size of the
evolved programs for different symbolic regression problems, and on the
computational complexity (based on how often the fitness function is
being executed). Use the following parameter configurations:
</p>
<ul>
<li>
<p> <b>Problems:</b> </p>
<p>
<table style="margin-left: 2em;">
<tr><td> Problem 1: </td><td> [$$] p_1 (x) = 5x^5 + 4x^4 + 3x^3 + 2x^2 + x [/$$] </td></tr>
<tr><td> Problem 2: </td><td> [$$] p_2 (x) = 4 \sin \left ( \frac {5 \pi}{4} x \right ) [/$$] </td></tr>
<tr><td> Problem 3: </td><td> [$$] p_3 (x) = G(x; -1.7, 0.5 ) + 7 G(x; 1.3, 0.8), [/$$] where [$] G [/$] is the Gaussian function:<span style="margin-left: 2em;"></span> [$] \displaystyle G(x; \mu, \sigma) = \frac {1}{ \sigma \sqrt{2 \pi} } \exp \left \{ -\frac{1}{2} \left ( \frac{ x - \mu }{ \sigma } \right )^2 \right \} [/$] </td></tr>
<tr><td> Problem 4: </td><td> [$$] p_4 (x) = ( \text{create your own function}) [/$$] </td></tr>
</table>
</p>
<p> Select <rb><i><span style="font-size: 1.4em;">two</span></i></rb> of
the above problems (i.e. two different [$] p_i [/$]'s) and create
symbolic regression data sets generated by associating the sixty-five
[$] x [/$] values between approximately [$] -\pi [/$] and [$] +\pi [/$]
in steps of [$] \pi / 32 [/$] (i.e. <span style="font-size: 0.8em;"> [$] -3.14159, -3.04341, -2.94524
~\dots~3.14159[/$]</span>) to sixty-five corresponding target values [$] t [/$]
computed via the formula [$] t = p_i (x) [/$].
</p>
</li>
<li>
<p> <b>Population size:</b> </p>
<hint> <b>500</b> vs. <b>2000</b> </hint>
</li>
<li>
<p> <b>Tournament size:</b> </p>
<hint> <b>2</b> vs. <b>5</b> </hint>
</li>
<li>
<p> Set the other parameters of the GP to the following values:
<ul>
<li> generations = 30 </li>
<li> crossover rate = 0.7 </li>
<li> mutation rate = 0.3 </li>
</ul>
</p>
</li>
</ul>
<blockquote style="background-color: #88ddff;"> Note that, analysing two
problems with two population sizes and two tournament sizes results in
<b>eight</b> combinations; consequently you need to run <b>eight</b>
different experiments.
</blockquote>
<p> Once the runs are completed, analyse the data provided by the system
and make sure you look carefully at the output produced in different
runs. Is the GP maximising or minimising the fitness? Then <rb
style="font-size: 1.1em;">summarize the results in form of a table</rb>,
reporting statistics obtained in different configurations when solving
your chosen problem. Please select statistics you think are most
appropriate to characterize the behaviour of the GP.
</p>
<p> Identify the parameter configuration that worked best for each problem
and then do a further set of 10 runs with each configuration chosen.
</p>
<p> You are now ready to start writing your report. In your report
explicitly address the following questions:
<ul> <bc>
<li> Did GP show some change in behaviour as the problem, population
size and tournament size were changed? Describe what happened.
</li>
<li> Can you provide a tentative explanation for the behaviours you
have observed?
</li>
<li> Was the behaviour represented by the statistics (e.g., averages)
in one configuration typical of all runs in that configuration or
did you see ample variations in behaviour across the runs done with
a configuration? If there were variations, can you explain why these
happen?
</li>
<li> Explain what criteria you used to identify the best parameter
configuration for the two problems.
</li>
<li> Were the statistics you got in the extra 10 runs performed with
your chosen optimal configuration for each problem consistent with
the statistics you obtained in the first batch of runs? If not, what
do you think happened and how can we find out what's the best
configuration?
</li>
<li> What conclusions can you draw about the suitability of GP to
solve the given problems?
</li>
<li> Feel free to produce and report additional plots or tables
(possibly including one or two typical runs) if this can provide
support for your explanations.
</li>
</bc> </ul>
</p>
<blockquote style="background-color: #ff88aa; margin-top: 2em;"> Note: This is a research focused
assignment. This means you are encouraged to do further analyses and
generate additional evidence that supports your conclusions. This will
allow you to create a strong and conclusive report.
</blockquote>
<br><br>
<!-- Marking Criteria -->
<h3 id="Part2_MarkingCriteria"> Marking Scheme for Part 2 </h3>
<p> Part 2 is worth a total of 50 marks. These will be divided as follows:
<table>
<tr><td> Structure and presentation of the report (including division into sections, formatting, diagrams, etc.) </td><td>[5 marks]</td></tr>
<tr><td> Accurate description of GP's behaviour in the runs performed (clear presentation of the results) </td><td>[10 marks]</td></tr>
<tr><td> Depth and correctness of the analysis of the runs and plausibility of the explanations provided (Discussion of the results) </td><td>[15 marks]</td></tr>
<tr><td> Concluding section, including summary of what has and hasn't been learnt and possible further explorations </td><td>[10 marks]</td></tr>
<tr><td> Supplementary analyses and further experiments </td><td>[10 marks]</td></tr>
</table>
<br><br>
<!-- Part 2 Notes -->
<h3 id="Part2_Notes_and_guidance" style="margin-top: 1em;"> Notes and guidance on Part 2 </h3>
<blockquote style="background-color: #ffaadd;">
<h3 style="margin-top: 1em;"> Notes and guidance on Part 2 (mini-project) </h3>
<ul>
<li>
<u><b>Own code or third-party GP toolbox?</b></u><br>
If you enjoy coding and want to get a deeper understanding of GP,
then you can modify the representation and crossover/mutation
operators of your Genetic Algorithm code and implement a GP system
for symbolic regression yourself. Alternatively, you can use the
third-party DEAP Python GP toolbox to run the experiments. Below I
provide some guidance on how to use the toolbox. Please check the
online documentation if you have further questions.<br> <span
style="color: #aa0000; font-weight: bold; font-size: 1.1em;">Note,
that the choice of library does not impact on the mark!</span>
<br><br>
</li>
<li>
<u><b>Guidance for users of the <tt>DEAP</tt> python package</b></u><br style="margin-bottom: 0.5em;">
The file <tt>CE310-GP-Mini-Project.py</tt> shows
an example of how to run the experiments. This can be found in the
Assessment Information card on Moodle, below the assignment
specification.
<p> Please familiarize yourself with the code
and start a set of runs (the default parameters are fine) and see
what happens. The code generates a log that records data related
to fitness, and the size of the evolved programs.
</p>
<p> The parameters can be adjusted in the <tt>PARAMETERS</tt>
section of the code, near the top of the module. The code also
provides you an example on how to create functions to create
target values. Feel free to change and adapt the code as needed
and implement the functions. For more information, please refer to
the <a style="color: #222266;"
href="https://deap.readthedocs.io/en/master/"><tt>DEAP</tt>
package documentation online</a>.
</p><br>
</li>
<li>
<u><b>What must be submitted to FASER?</b></u><br>
Submit your code and the report <a style="color: #222266;"
href="#Part1_Notes_and_guidance">as discussed previously for Part
1</a>. For the Part 2 report, please be short and concise, and <b>do
not exceed 600 words</b> (not including Table or Table/Figure
captions). <br><br>
</li>
</ul>
</blockquote>
<br>
<flex><img src="data:image/png;base64,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" alt="[Floral Divider]"></flex>
<br>
<blockquote id="ProblemsEtc" style="background-color: #ff2222;">
<p>
<ul>
<li>
<u><b>Problems:</b></u><br>
If any general problems arise in the assignment, please start a
thread on the <a
href="https://moodle.essex.ac.uk/mod/forum/view.php?id=1203776">Help
Forum</a> or <a
href="https://moodle.essex.ac.uk/mod/moodleoverflow/view.php?id=1203777">MoodleOverflow</a>
on Moodle as appropriate. For specific/personal inquiries,
please use the academic support hour or <a
href="mailto:[email protected]">email me</a>.
</li>
<li style="margin-top: 2em;">
<u><b>Late Submissions, Extenuating Circumstances, Academic Offenses, etc:</b></u><br>
Please refer to the Students’ Handbook for details regarding the
Departmental policy for the above topics. Note that the module
supervisors are <span style="color: #ffdddd; font-size: 1.2em;
font-weight: bold;">not</span> part of this process at all; it is
all handled externally by the School Office and relevant committees.
For any questions regarding any of the above topics, please email
the <a href="mailto:[email protected]">School Office</a>
directly.
</li>
</ul>
</p>
</blockquote>
<div style="text-align: center; width: 100%; font-size: 1.5em; margin-top: 2em;">
<b>•·················•·················•</b>
<p> That is all. Good luck! :)</p>
</div>
<!-- ========================== -->
<!-- End of assessment document -->
<!-- ========================== -->
</centralcontainer>
</body>