Datasets:
Tasks:
Text Generation
Modalities:
Text
Sub-tasks:
language-modeling
Languages:
English
Size:
100K - 1M
License:
File size: 15,087 Bytes
4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 fc5e983 4365a98 |
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 |
/-
Copyright (c) 2020 Fox Thomson. All rights reserved.
Released under Apache 2.0 license as described in the file LICENSE.
Authors: Fox Thomson, Markus Himmel
-/
import data.nat.bitwise
import set_theory.game.birthday
import set_theory.game.impartial
/-!
# Nim and the Sprague-Grundy theorem
This file contains the definition for nim for any ordinal `o`. In the game of `nim o₁` both players
may move to `nim o₂` for any `o₂ < o₁`.
We also define a Grundy value for an impartial game `G` and prove the Sprague-Grundy theorem, that
`G` is equivalent to `nim (grundy_value G)`.
Finally, we compute the sum of finite Grundy numbers: if `G` and `H` have Grundy values `n` and `m`,
where `n` and `m` are natural numbers, then `G + H` has the Grundy value `n xor m`.
## Implementation details
The pen-and-paper definition of nim defines the possible moves of `nim o` to be `set.Iio o`.
However, this definition does not work for us because it would make the type of nim
`ordinal.{u} → pgame.{u + 1}`, which would make it impossible for us to state the Sprague-Grundy
theorem, since that requires the type of `nim` to be `ordinal.{u} → pgame.{u}`. For this reason, we
instead use `o.out.α` for the possible moves. You can use `to_left_moves_nim` and
`to_right_moves_nim` to convert an ordinal less than `o` into a left or right move of `nim o`, and
vice versa.
-/
noncomputable theory
universe u
open_locale pgame
namespace pgame
/-- The definition of single-heap nim, which can be viewed as a pile of stones where each player can
take a positive number of stones from it on their turn. -/
-- Uses `noncomputable!` to avoid `rec_fn_macro only allowed in meta definitions` VM error
noncomputable! def nim : ordinal.{u} → pgame.{u}
| o₁ :=
let f := λ o₂,
have ordinal.typein o₁.out.r o₂ < o₁ := ordinal.typein_lt_self o₂,
nim (ordinal.typein o₁.out.r o₂)
in ⟨o₁.out.α, o₁.out.α, f, f⟩
using_well_founded { dec_tac := tactic.assumption }
open ordinal
lemma nim_def (o : ordinal) : nim o = pgame.mk o.out.α o.out.α
(λ o₂, nim (ordinal.typein (<) o₂))
(λ o₂, nim (ordinal.typein (<) o₂)) :=
by { rw nim, refl }
lemma left_moves_nim (o : ordinal) : (nim o).left_moves = o.out.α :=
by { rw nim_def, refl }
lemma right_moves_nim (o : ordinal) : (nim o).right_moves = o.out.α :=
by { rw nim_def, refl }
lemma move_left_nim_heq (o : ordinal) : (nim o).move_left == λ i : o.out.α, nim (typein (<) i) :=
by { rw nim_def, refl }
lemma move_right_nim_heq (o : ordinal) : (nim o).move_right == λ i : o.out.α, nim (typein (<) i) :=
by { rw nim_def, refl }
/-- Turns an ordinal less than `o` into a left move for `nim o` and viceversa. -/
noncomputable def to_left_moves_nim {o : ordinal} : set.Iio o ≃ (nim o).left_moves :=
(enum_iso_out o).to_equiv.trans (equiv.cast (left_moves_nim o).symm)
/-- Turns an ordinal less than `o` into a right move for `nim o` and viceversa. -/
noncomputable def to_right_moves_nim {o : ordinal} : set.Iio o ≃ (nim o).right_moves :=
(enum_iso_out o).to_equiv.trans (equiv.cast (right_moves_nim o).symm)
@[simp] theorem to_left_moves_nim_symm_lt {o : ordinal} (i : (nim o).left_moves) :
↑(to_left_moves_nim.symm i) < o :=
(to_left_moves_nim.symm i).prop
@[simp] theorem to_right_moves_nim_symm_lt {o : ordinal} (i : (nim o).right_moves) :
↑(to_right_moves_nim.symm i) < o :=
(to_right_moves_nim.symm i).prop
@[simp] lemma move_left_nim' {o : ordinal.{u}} (i) :
(nim o).move_left i = nim (to_left_moves_nim.symm i).val :=
(congr_heq (move_left_nim_heq o).symm (cast_heq _ i)).symm
lemma move_left_nim {o : ordinal} (i) :
(nim o).move_left (to_left_moves_nim i) = nim i :=
by simp
@[simp] lemma move_right_nim' {o : ordinal} (i) :
(nim o).move_right i = nim (to_right_moves_nim.symm i).val :=
(congr_heq (move_right_nim_heq o).symm (cast_heq _ i)).symm
lemma move_right_nim {o : ordinal} (i) :
(nim o).move_right (to_right_moves_nim i) = nim i :=
by simp
instance is_empty_nim_zero_left_moves : is_empty (nim 0).left_moves :=
by { rw nim_def, exact ordinal.is_empty_out_zero }
instance is_empty_nim_zero_right_moves : is_empty (nim 0).right_moves :=
by { rw nim_def, exact ordinal.is_empty_out_zero }
/-- `nim 0` has exactly the same moves as `0`. -/
def nim_zero_relabelling : nim 0 ≡r 0 := relabelling.is_empty _
theorem nim_zero_equiv : nim 0 ≈ 0 := equiv.is_empty _
noncomputable instance unique_nim_one_left_moves : unique (nim 1).left_moves :=
(equiv.cast $ left_moves_nim 1).unique
noncomputable instance unique_nim_one_right_moves : unique (nim 1).right_moves :=
(equiv.cast $ right_moves_nim 1).unique
@[simp] theorem default_nim_one_left_moves_eq :
(default : (nim 1).left_moves) = @to_left_moves_nim 1 ⟨0, ordinal.zero_lt_one⟩ :=
rfl
@[simp] theorem default_nim_one_right_moves_eq :
(default : (nim 1).right_moves) = @to_right_moves_nim 1 ⟨0, ordinal.zero_lt_one⟩ :=
rfl
@[simp] theorem to_left_moves_nim_one_symm (i) :
(@to_left_moves_nim 1).symm i = ⟨0, ordinal.zero_lt_one⟩ :=
by simp
@[simp] theorem to_right_moves_nim_one_symm (i) :
(@to_right_moves_nim 1).symm i = ⟨0, ordinal.zero_lt_one⟩ :=
by simp
theorem nim_one_move_left (x) : (nim 1).move_left x = nim 0 :=
by simp
theorem nim_one_move_right (x) : (nim 1).move_right x = nim 0 :=
by simp
/-- `nim 1` has exactly the same moves as `star`. -/
def nim_one_relabelling : nim 1 ≡r star :=
begin
rw nim_def,
refine ⟨_, _, λ i, _, λ j, _⟩,
any_goals { dsimp, apply equiv.equiv_of_unique },
all_goals { simp, exact nim_zero_relabelling }
end
theorem nim_one_equiv : nim 1 ≈ star := nim_one_relabelling.equiv
@[simp] lemma nim_birthday (o : ordinal) : (nim o).birthday = o :=
begin
induction o using ordinal.induction with o IH,
rw [nim_def, birthday_def],
dsimp,
rw max_eq_right le_rfl,
convert lsub_typein o,
exact funext (λ i, IH _ (typein_lt_self i))
end
@[simp] lemma neg_nim (o : ordinal) : -nim o = nim o :=
begin
induction o using ordinal.induction with o IH,
rw nim_def, dsimp; congr;
funext i;
exact IH _ (ordinal.typein_lt_self i)
end
instance nim_impartial (o : ordinal) : impartial (nim o) :=
begin
induction o using ordinal.induction with o IH,
rw [impartial_def, neg_nim],
refine ⟨equiv_rfl, λ i, _, λ i, _⟩;
simpa using IH _ (typein_lt_self _)
end
lemma exists_ordinal_move_left_eq {o : ordinal} (i) : ∃ o' < o, (nim o).move_left i = nim o' :=
⟨_, typein_lt_self _, move_left_nim' i⟩
lemma exists_move_left_eq {o o' : ordinal} (h : o' < o) : ∃ i, (nim o).move_left i = nim o' :=
⟨to_left_moves_nim ⟨o', h⟩, by simp⟩
lemma nim_fuzzy_zero_of_ne_zero {o : ordinal} (ho : o ≠ 0) : nim o ∥ 0 :=
begin
rw [impartial.fuzzy_zero_iff_lf, nim_def, lf_zero_le],
rw ←ordinal.pos_iff_ne_zero at ho,
exact ⟨(ordinal.principal_seg_out ho).top, by simp⟩
end
@[simp] lemma nim_add_equiv_zero_iff (o₁ o₂ : ordinal) : nim o₁ + nim o₂ ≈ 0 ↔ o₁ = o₂ :=
begin
split,
{ contrapose,
intro h,
rw [impartial.not_equiv_zero_iff],
wlog h' : o₁ ≤ o₂ using [o₁ o₂, o₂ o₁],
{ exact le_total o₁ o₂ },
{ have h : o₁ < o₂ := lt_of_le_of_ne h' h,
rw [impartial.fuzzy_zero_iff_gf, zero_lf_le, nim_def o₂],
refine ⟨to_left_moves_add (sum.inr _), _⟩,
{ exact (ordinal.principal_seg_out h).top },
{ simpa using (impartial.add_self (nim o₁)).2 } },
{ exact (fuzzy_congr_left add_comm_equiv).1 (this (ne.symm h)) } },
{ rintro rfl,
exact impartial.add_self (nim o₁) }
end
@[simp] lemma nim_add_fuzzy_zero_iff {o₁ o₂ : ordinal} : nim o₁ + nim o₂ ∥ 0 ↔ o₁ ≠ o₂ :=
by rw [iff_not_comm, impartial.not_fuzzy_zero_iff, nim_add_equiv_zero_iff]
@[simp] lemma nim_equiv_iff_eq {o₁ o₂ : ordinal} : nim o₁ ≈ nim o₂ ↔ o₁ = o₂ :=
by rw [impartial.equiv_iff_add_equiv_zero, nim_add_equiv_zero_iff]
/-- The Grundy value of an impartial game, the ordinal which corresponds to the game of nim that the
game is equivalent to -/
noncomputable def grundy_value : Π (G : pgame.{u}), ordinal.{u}
| G := ordinal.mex.{u u} (λ i, grundy_value (G.move_left i))
using_well_founded { dec_tac := pgame_wf_tac }
lemma grundy_value_eq_mex_left (G : pgame) :
grundy_value G = ordinal.mex.{u u} (λ i, grundy_value (G.move_left i)) :=
by rw grundy_value
/-- The Sprague-Grundy theorem which states that every impartial game is equivalent to a game of
nim, namely the game of nim corresponding to the games Grundy value -/
theorem equiv_nim_grundy_value : ∀ (G : pgame.{u}) [G.impartial], G ≈ nim (grundy_value G)
| G :=
begin
introI hG,
rw [impartial.equiv_iff_add_equiv_zero, ←impartial.forall_left_moves_fuzzy_iff_equiv_zero],
intro i,
apply left_moves_add_cases i,
{ intro i₁,
rw add_move_left_inl,
apply (fuzzy_congr_left (add_congr_left (equiv_nim_grundy_value (G.move_left i₁)).symm)).1,
rw nim_add_fuzzy_zero_iff,
intro heq,
rw [eq_comm, grundy_value_eq_mex_left G] at heq,
have h := ordinal.ne_mex _,
rw heq at h,
exact (h i₁).irrefl },
{ intro i₂,
rw [add_move_left_inr, ←impartial.exists_left_move_equiv_iff_fuzzy_zero],
revert i₂,
rw nim_def,
intro i₂,
have h' : ∃ i : G.left_moves, (grundy_value (G.move_left i)) =
ordinal.typein (quotient.out (grundy_value G)).r i₂,
{ revert i₂,
rw grundy_value_eq_mex_left,
intros i₂,
have hnotin : _ ∉ _ := λ hin, (le_not_le_of_lt (ordinal.typein_lt_self i₂)).2 (cInf_le' hin),
simpa using hnotin},
cases h' with i hi,
use to_left_moves_add (sum.inl i),
rw [add_move_left_inl, move_left_mk],
apply (add_congr_left (equiv_nim_grundy_value (G.move_left i))).trans,
simpa only [hi] using impartial.add_self (nim (grundy_value (G.move_left i))) }
end
using_well_founded { dec_tac := pgame_wf_tac }
lemma grundy_value_eq_iff_equiv_nim {G : pgame} [G.impartial] {o : ordinal} :
grundy_value G = o ↔ G ≈ nim o :=
⟨by { rintro rfl, exact equiv_nim_grundy_value G },
by { intro h, rw ←nim_equiv_iff_eq, exact (equiv_nim_grundy_value G).symm.trans h }⟩
@[simp] lemma nim_grundy_value (o : ordinal.{u}) : grundy_value (nim o) = o :=
grundy_value_eq_iff_equiv_nim.2 pgame.equiv_rfl
lemma grundy_value_eq_iff_equiv (G H : pgame) [G.impartial] [H.impartial] :
grundy_value G = grundy_value H ↔ G ≈ H :=
grundy_value_eq_iff_equiv_nim.trans (equiv_congr_left.1 (equiv_nim_grundy_value H) _).symm
@[simp] lemma grundy_value_zero : grundy_value 0 = 0 :=
grundy_value_eq_iff_equiv_nim.2 nim_zero_equiv.symm
lemma grundy_value_iff_equiv_zero (G : pgame) [G.impartial] : grundy_value G = 0 ↔ G ≈ 0 :=
by rw [←grundy_value_eq_iff_equiv, grundy_value_zero]
@[simp] lemma grundy_value_star : grundy_value star = 1 :=
grundy_value_eq_iff_equiv_nim.2 nim_one_equiv.symm
@[simp] lemma grundy_value_neg (G : pgame) [G.impartial] : grundy_value (-G) = grundy_value G :=
by rw [grundy_value_eq_iff_equiv_nim, neg_equiv_iff, neg_nim, ←grundy_value_eq_iff_equiv_nim]
lemma grundy_value_eq_mex_right : ∀ (G : pgame) [G.impartial],
grundy_value G = ordinal.mex.{u u} (λ i, grundy_value (G.move_right i))
| ⟨l, r, L, R⟩ := begin
introI H,
rw [←grundy_value_neg, grundy_value_eq_mex_left],
congr,
ext i,
haveI : (R i).impartial := @impartial.move_right_impartial ⟨l, r, L, R⟩ _ i,
apply grundy_value_neg
end
@[simp] lemma grundy_value_nim_add_nim (n m : ℕ) :
grundy_value (nim.{u} n + nim.{u} m) = nat.lxor n m :=
begin
induction n using nat.strong_induction_on with n hn generalizing m,
induction m using nat.strong_induction_on with m hm,
rw [grundy_value_eq_mex_left],
-- We want to show that `n xor m` is the smallest unreachable Grundy value. We will do this in two
-- steps:
-- h₀: `n xor m` is not a reachable grundy number.
-- h₁: every Grundy number strictly smaller than `n xor m` is reachable.
have h₀ : ∀ i, grundy_value ((nim n + nim m).move_left i) ≠ (nat.lxor n m : ordinal),
{ -- To show that `n xor m` is unreachable, we show that every move produces a Grundy number
-- different from `n xor m`.
intro i,
-- The move operates either on the left pile or on the right pile.
apply left_moves_add_cases i,
all_goals
{ -- One of the piles is reduced to `k` stones, with `k < n` or `k < m`.
intro a,
obtain ⟨ok, hk, hk'⟩ := exists_ordinal_move_left_eq a,
obtain ⟨k, rfl⟩ := ordinal.lt_omega.1 (lt_trans hk (ordinal.nat_lt_omega _)),
replace hk := ordinal.nat_cast_lt.1 hk,
-- Thus, the problem is reduced to computing the Grundy value of `nim n + nim k` or
-- `nim k + nim m`, both of which can be dealt with using an inductive hypothesis.
simp only [hk', add_move_left_inl, add_move_left_inr, id],
rw hn _ hk <|> rw hm _ hk,
-- But of course xor is injective, so if we change one of the arguments, we will not get the
-- same value again.
intro h,
rw ordinal.nat_cast_inj at h,
try { rw [nat.lxor_comm n k, nat.lxor_comm n m] at h },
exact hk.ne (nat.lxor_left_injective h) } },
have h₁ : ∀ (u : ordinal), u < nat.lxor n m →
u ∈ set.range (λ i, grundy_value ((nim n + nim m).move_left i)),
{ -- Take any natural number `u` less than `n xor m`.
intros ou hu,
obtain ⟨u, rfl⟩ := ordinal.lt_omega.1 (lt_trans hu (ordinal.nat_lt_omega _)),
replace hu := ordinal.nat_cast_lt.1 hu,
-- Our goal is to produce a move that gives the Grundy value `u`.
rw set.mem_range,
-- By a lemma about xor, either `u xor m < n` or `u xor n < m`.
cases nat.lt_lxor_cases hu with h h,
-- Therefore, we can play the corresponding move, and by the inductive hypothesis the new state
-- is `(u xor m) xor m = u` or `n xor (u xor n) = u` as required.
{ obtain ⟨i, hi⟩ := exists_move_left_eq (ordinal.nat_cast_lt.2 h),
refine ⟨to_left_moves_add (sum.inl i), _⟩,
simp only [hi, add_move_left_inl],
rw [hn _ h, nat.lxor_assoc, nat.lxor_self, nat.lxor_zero] },
{ obtain ⟨i, hi⟩ := exists_move_left_eq (ordinal.nat_cast_lt.2 h),
refine ⟨to_left_moves_add (sum.inr i), _⟩,
simp only [hi, add_move_left_inr],
rw [hm _ h, nat.lxor_comm, nat.lxor_assoc, nat.lxor_self, nat.lxor_zero] } },
-- We are done!
apply (ordinal.mex_le_of_ne.{u u} h₀).antisymm,
contrapose! h₁,
exact ⟨_, ⟨h₁, ordinal.mex_not_mem_range _⟩⟩,
end
lemma nim_add_nim_equiv {n m : ℕ} : nim n + nim m ≈ nim (nat.lxor n m) :=
by rw [←grundy_value_eq_iff_equiv_nim, grundy_value_nim_add_nim]
lemma grundy_value_add (G H : pgame) [G.impartial] [H.impartial] {n m : ℕ} (hG : grundy_value G = n)
(hH : grundy_value H = m) : grundy_value (G + H) = nat.lxor n m :=
begin
rw [←nim_grundy_value (nat.lxor n m), grundy_value_eq_iff_equiv],
refine equiv.trans _ nim_add_nim_equiv,
convert add_congr (equiv_nim_grundy_value G) (equiv_nim_grundy_value H);
simp only [hG, hH]
end
end pgame
|