T05: publish Benchmarking & Parameter Tuning slides

Mirror the T05 deck (notes stripped) and link it on the slides index.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: d-krupke

slides/T05/T05-benchmarking-tuning.qmd

@@ -0,0 +1,47 @@
+---
+title: "Benchmarking & Parameter Tuning"
+subtitle: "Algorithm Engineering — T05"
+author: Dr. Dominik Krupke
+format:
+  revealjs:
+    slide-number: true
+    width: 1600
+    height: 900
+    chalkboard:
+      buttons: false
+    preview-links: auto
+    css: styles.css
+    footer: "Algorithm Engineering SS 2026 — T05 Benchmarking & Parameter Tuning"
+    progress: true
+    transition: none
+    background-transition: none
+    theme: dark
+    highlight-style: github
+    code-overflow: wrap
+bibliography: references.bib
+---
+
+<!-- ============================================================
+     T05 — Benchmarking & Parameter Tuning
+     Two halves:
+       (A) Benchmarking            _00 .. _02
+       (B) Log reading + tuning    _03 .. _06
+     Primary sources (cpsat-primer, ~/Repositories/cpsat-primer):
+       - chapters/benchmarking.md           -> _01, _02
+       - chapters/understanding_the_log.md  -> _03
+       - chapters/parameters.md             -> _03 (subsolver contribution), _04
+     ============================================================ -->
+
+{{< include _00-intro.qmd >}}
+{{< include _01-benchmarking-pitfalls.qmd >}}
+{{< include _02-benchmarking-visualization.qmd >}}
+{{< include _02b-study-design.qmd >}}
+{{< include _03-reading-the-log.qmd >}}
+{{< include _04-tuning-search.qmd >}}
+{{< include _05-optuna-tutorial.qmd >}}
+{{< include _06-wrap-up.qmd >}}
+
+## References
+
+::: {#refs}
+:::

slides/T05/_00-intro.qmd

@@ -0,0 +1,87 @@
+<!-- Sources for Intro:
+     - Framing: README.md key message; benchmarking.md intro (§"Benchmarking your Model")
+     - ASSET TODO: assets/symbol_benchmarking_tuning.png (section-divider background,
+       dark/atmospheric, match L01/T01 symbol style)
+-->
+
+# Benchmarking & Parameter Tuning {background-image="assets/symbol_benchmarking_tuning.png" background-opacity="0.3" background-size="cover" background-color="#2d4059"}
+
+You changed the model and it got faster on your three test instances. Did it actually get better?
+
+## What this tutorial is about
+
+:::: {.columns .viz-row}
+::: {.column width="33%"}
+![](assets/cactus_plot.png)
+:::
+::: {.column width="33%"}
+::: {.loglet}
+```
+Starting CP-SAT solver v9.15.6755
+Setting number of workers to 24
+
+Initial optimization model 'multiple_knapsack':
+#Variables: 560 (#bools: 560)
+#kLinearN: 144 (#terms: 1'120)
+
+Starting presolve at 0.00s
+[Symmetry] #generators: 7, average support size: 8
+presolved: 560 vars, 144 constraints
+
+Starting search at 0.01s with 24 workers
+#1     0.01s best:1891  next:[1892,3797] no_lp
+#5     0.01s best:1939  next:[1940,2271] core
+#10    0.01s best:2226  next:[2227,2271] default_lp
+#13    0.02s best:2266  next:[2267,2271] reduced_costs
+#Bound 0.05s best:2267  next:[2268,2270] probing_max_lp
+#15    0.08s best:2269  next:[2270,2270] default_lp
+#16    0.18s best:2270  next:[]          graph_var_lns
+
+Solutions (7)             Num   Rank
+                'no_lp':    3  [1,7]
+        'quick_restart':    1  [3,3]
+  'quick_restart_no_lp':    3  [2,5]
+
+CpSolverResponse summary:
+status: OPTIMAL
+objective: 2270
+best_bound: 2270
+gap_integral: 0.44563
+```
+:::
+:::
+::: {.column width="33%"}
+![](assets/surrogate_acquisition_5.png)
+:::
+::::
+
+:::: {.columns .col-titles}
+::: {.column width="33%"}
+### Benchmarking
+:::
+::: {.column width="33%"}
+### Reading the log
+:::
+::: {.column width="33%"}
+### Parameter tuning
+:::
+::::
+
+:::: {.columns .bullet-row .smaller}
+::: {.column width="33%"}
+- Is a "win" real, or three lucky instances?
+- Which metric and plot fit the goal?
+- Timeouts, variance, and the right instance set.
+:::
+::: {.column width="33%"}
+- What did the solver actually do?
+- Which strategies moved the bounds?
+- Where is the easy tuning win?
+:::
+::: {.column width="33%"}
+- Tuning as a search problem.
+- Grid, random, then model-based.
+- A hands-on recipe to run it.
+:::
+::::
+