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
| #include <iostream>
#include <vector>
#include <cmath>
#include <algorithm>
#include <cstdint>
#include <random>
#include <iomanip>
#include <bits/stdc++.h>
struct Timer {
std::chrono::high_resolution_clock::time_point st;
Timer() { st = now(); }
std::chrono::high_resolution_clock::time_point now() { return std::chrono::high_resolution_clock::now(); }
std::chrono::milliseconds::rep span() {
auto ed = now();
return std::chrono::duration_cast<std::chrono::milliseconds>(ed - st).count();
}
};
using u32 = std::uint_fast32_t;
using i64 = std::int_fast64_t;
constexpr int K = 26;
int D;
std::vector<i64> C;
std::vector<std::vector<i64>> S;
using Score = i64;
struct Individual {
std::vector<int> vec;
Individual() = default;
explicit Individual(std::vector<int> vec): vec(std::move(vec)) {}
Individual crossover(std::mt19937& mt, const Individual& y) const {
Individual mx;
Score score = -1e9;
int now = vec.size() / 2;
{
std::vector<int> next(vec.size());
std::copy(vec.begin(), vec.begin() + vec.size() / 2, next.begin());
std::copy(y.vec.begin() + vec.size() / 2, y.vec.end(), next.begin() + vec.size() / 2);
Individual child(std::move(next));
score = child.evaluate_state();
mx = std::move(child);
}
constexpr int Count = 30;
for(int q = 0; q < Count; q++) {
int DIFF = 15;
int at = std::min(D - 1, std::max(1, std::uniform_int_distribution<>(-DIFF, DIFF - 1)(mt) + now));
if(at >= 0) at++;
std::vector<int> next(vec.size());
std::copy(vec.begin(), vec.begin() + at, next.begin());
std::copy(y.vec.begin() + at, y.vec.end(), next.begin() + at);
Individual child(std::move(next));
Score next_score = child.evaluate_state();
if(score <= next_score || std::bernoulli_distribution(std::exp(-(score - next_score) / 4e2))(mt)) {
score = next_score;
mx = std::move(child);
now = at;
}
}
return mx;
}
void mutation(std::mt19937& mt) {
std::bernoulli_distribution swap_dist(0.005);
for(int i = 1; i < vec.size(); i++) {
if(swap_dist(mt)) {
std::uniform_int_distribution<> at_dist(std::max(0, i - 15), i - 1);
int j = at_dist(mt);
std::swap(vec[i], vec[j]);
}
}
for(int i = 0; i < vec.size(); i++) {
if(swap_dist(mt)) {
vec[i] = std::uniform_int_distribution<>(0, K - 1)(mt);
}
}
}
Score evaluate_state() const {
i64 sum = 0;
std::vector<i64> before(K, -1);
for(int i = 0; i < D; i++) {
int v = vec[i];
sum += S[i][v];
i64 d = i - before[v];
sum -= C[v] * d * (d - 1) / 2;
before[v] = i;
}
for(int v = 0; v < K; v++) {
i64 d = D - before[v];
sum -= C[v] * d * (d - 1) / 2;
}
return sum;
}
};
void local_search(std::mt19937& mt, Individual& x, Score& score) {
constexpr int Count = 50;
constexpr double T = 5e2;
for(int q = 0; q < Count; q++) {
if(std::bernoulli_distribution(0.5)(mt)) {
int i = std::uniform_int_distribution<>(0, x.vec.size() - 1)(mt);
int j = std::uniform_int_distribution<>(std::max(0, i - 13), std::min(D - 2, i + 12))(mt);
if(j >= i) j++;
std::swap(x.vec[i], x.vec[j]);
Score next = x.evaluate_state();
if(score <= next || std::bernoulli_distribution(std::exp(-(score - next) / T))(mt)) {
score = next;
}
else {
std::swap(x.vec[i], x.vec[j]);
}
}
else {
int i = std::uniform_int_distribution<>(0, x.vec.size() - 1)(mt);
int before = x.vec[i];
x.vec[i] = std::uniform_int_distribution<>(0, K - 2)(mt);
if(before <= x.vec[i]) x.vec[i]++;
Score next = x.evaluate_state();
if(score <= next || std::bernoulli_distribution(std::exp(-(score - next) / T))(mt)) {
score = next;
}
else {
x.vec[i] = before;
}
}
}
}
struct Generation {
constexpr static int Count = 12;
constexpr static int Elite = 4;
constexpr static int NewInd = 3;
std::vector<std::pair<Individual, Score>> inds;
void init(std::mt19937& mt) {
for(int i = 0; i < Count; i++) {
std::vector<int> init(D);
for(int d = 0; d < D; d++) {
init[d] = std::uniform_int_distribution<>(0, K - 1)(mt);
}
Individual x(std::move(init));
Score score = x.evaluate_state();
inds.emplace_back(std::move(x), std::move(score));
}
}
Generation next_gen(std::mt19937& mt) const {
Generation next;
std::vector<int> idx(Count);
std::iota(idx.begin(), idx.end(), 0);
// elitism
std::nth_element(idx.begin(), idx.begin() + Elite - 1, idx.end(), [&](int i, int j) { return inds[i].second > inds[j].second; });
for(int i = 0; i < Elite; i++) {
next.inds.push_back(inds[idx[i]]);
}
// selection & crossover by roulette
Score max_score = std::max_element(inds.begin(), inds.end(), [](auto& a, auto& b) { return a.second > b.second; })->second;
std::vector<double> pie(inds.size());
for(int i = 0; i < inds.size(); i++) {
pie[i] = inds[i].second - max_score;
if(i > 0) {
pie[i] += pie[i - 1];
}
}
std::uniform_real_distribution<> dice(0, pie.back());
for(int i = Elite; i < Count - NewInd; i++) {
int x = std::lower_bound(pie.begin(), pie.end(), dice(mt), [](auto& a, double v) { return a < v; }) - pie.begin();
double diff = pie[x] - (x == 0 ? 0 : pie[x - 1]);
int y = std::lower_bound(pie.begin(), pie.end(), std::uniform_real_distribution<>(0, pie.back() - diff)(mt),
[&](auto& a, double v) { return (a >= pie[x] ? a - diff : a ) < v; }
) - pie.begin();
Individual child = inds[x].first.crossover(mt, inds[y].first);
child.mutation(mt);
Score score = child.evaluate_state();
next.inds.emplace_back(std::move(child), std::move(score));
}
for(int i = Count - NewInd; i < Count; i++) {
std::vector<int> init(D);
for(int d = 0; d < D; d++) {
init[d] = std::uniform_int_distribution<>(0, K - 1)(mt);
}
Individual x(std::move(init));
/*
Individual x(next.inds[i - Count + NewInd].first);
x.mutation(mt);
x.mutation(mt);
*/
Score score = x.evaluate_state();
next.inds.emplace_back(std::move(x), std::move(score));
}
for(auto& [x, score]: next.inds) {
local_search(mt, x, score);
}
return next;
}
};
int main() {
std::mt19937 mt;
//const int Century = 100000;
std::cin >> D;
C.resize(K);
for(int i = 0; i < K; i++) {
std::cin >> C[i];
}
S.resize(D, std::vector<i64>(K));
for(int i = 0; i < D; i++) {
for(int j = 0; j < K; j++) {
std::cin >> S[i][j];
}
}
Generation gen;
gen.init(mt);
Timer timer;
//for(int q = 0; q < Century; q++) {
int q = 0;
Individual MAX;
Score score = -1e9;
while(timer.span() < 1980) {
gen = gen.next_gen(mt);
q++;
if(q % 100 == 0) {
auto& NEXT = *std::max_element(gen.inds.begin(), gen.inds.end(), [](const auto& a, const auto& b) { return a.second < b.second; });
if(score < NEXT.second) {
MAX = NEXT.first;
score = NEXT.second;
}
std::cerr << q << " " << score << std::endl;
}
//std::cerr << std::fixed << std::setprecision(10) << q++ << "\t" << MAX.second << std::endl;
}
std::cerr << timer.span() << std::endl;
std::cerr << score << std::endl;
for(int i = 0; i < D; i++) {
std::cout << MAX.vec[i] + 1 << "\n";
}
}
|