barrett/barrettruth.com
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

feat(cp-log): 871

Browse source
This commit is contained in:
Barrett Ruth 2025-03-28 13:00:14 -04:00
parent 77807707da
commit a727ce90cd
1 changed files with 89 additions and 0 deletions
Download patch file Download diff file Expand all files Collapse all files

89
posts/algorithms/cp-log.html
View file

@ -35,6 +35,95 @@
<h1 class="post-title">Competitive Programming Log</h1>
</header>
<article class="post-article">
<h2>
<a href="https://codeforces.com/contest/1892/" target="_blank"
>871 (div. 4)</a
>&mdash;28/3/2025
</h2>
<div>
Div. 4 to practice implemenation skills + mathematical observations
thanks to
<a
href="http://www.gang.umass.edu/~franz/Paul_Zeitz_The_Art_and_Craft_of_Problem_SolvingBookosorg.pdf"
target="_blank"
>Paul Zeitz</a
>. From now on, I will only note useful problems.
</div>
<ul>
<li>
B: typo, costing a few minutes. Go slower. Declare variables.
Think consistently through approach.
</li>
<li>
C: paused an found a general implementation as Zeitz advises
(starting with a general, non-mathematical solution: "The solution
is the earliest to get either both 1s at once or each over two
strings"). <b>I still rushed</b>, incorrectly computing the result
as
<code>min(first, second)</code>
instead of
<code>first + second</code>.
</li>
<li>
D: I looked up the recurrence relation \(T(n)=T(n/3)+T(2n/3)\) to
ensure it was sub-linear time.
<blockquote style="font-style: italic">
Gain the mathematical skills to both analyze and derive (i.e. if
you forget) recurrence relations.
</blockquote>
</li>
<li>
E: spent <i>forever</i> fanagling with my Disjoint Set Union
template. I made countless errors including:
<ol>
<li>
Not inserting nodes into the data structure before joining
them
</li>
<li>
Joining zero/non-zero nodes, corrupting the component graph
</li>
<li>
Erroneously updating component sums by double counting (i.e.
using a mutated value).
</li>
</ol>
<blockquote>
Understand data structures before you use them. Despite what
Colin Galen says, black-boxed knowledge often isn't enough.
Improve knowledge of Fenwick Trees, Segment Trees, and Union
Find.
</blockquote>
</li>
<li>
F: elegant approach and reduction, eliminating the need for a
graph traversal.
</li>
<li>
G: <i>intentionally</i> chose the mathematical/indexing method to
improve my skills. I'd also like to use
<code>{upper,lower}_bound</code> more often with a lambda than
manually code binary search.
<b
>Once again, I practiced in a fundamentally flawed way, choosing
an approach I knew would be error prone and difficult</b
>.
<blockquote>
Don't be scared if you initilly can't find an easy
implementation.
<i>Reconceptualize, visualize, and reframe</i> until something
comes up. Sometimes, this takes too long&mdash;however, my
threshold for that realization is MUCH too low.
<b
>Spend a longer time developing/considering approaches even if
you already know a feasible solution</b
>. Here, I spent ~5 minutes developing a solution, then 20
minutes coding it. Instead, another 5 minute allocation of time
could lead me to the prefix sum solution, likely saving ~10
minutes.
</blockquote>
</li>
</ul>
<h2>
<a href="https://codeforces.com/contest/1872/" target="_blank"
>895 (div. 3)</a