結果

問題 No.5018 Let's Make a Best-seller Book
ユーザー Kiri8128
提出日時 2023-10-02 16:43:02
言語 PyPy3
(7.3.15)
結果
AC  
実行時間 181 ms / 400 ms
コード長 17,997 bytes
コンパイル時間 364 ms
コンパイル使用メモリ 86,972 KB
実行使用メモリ 94,652 KB
スコア 174,383
平均クエリ数 52.00
最終ジャッジ日時 2023-10-02 16:43:28
合計ジャッジ時間 22,706 ms
ジャッジサーバーID
（参考情報） 		judge14 / judge15
純コード判定しない問題か言語
このコードへのチャレンジ
（要ログイン）
ファイルパターン 結果
other AC * 100
権限があれば一括ダウンロードができます

ソースコード

diff #
プレゼンテーションモードにする

from random import gauss, random, randrange
try:
    LOCAL
except NameError:
    LOCAL = 0
if LOCAL:
    import optuna
    DEBUG = 1
    print("LOOP ?")
    LOOP = int(input())
    OP = 1 if LOOP == 0 else 0
else:
    DEBUG = 0
    LOOP = 1
    OP = 0
def estimate(i, p):
    l, r = int(MI[i] * 100 + 0.999), int(MA[i] * 100) + 1
    if MA[i] - MI[i] < 0.01 or l >= r - 1:
        return MI[i] + (MA[i] - MI[i]) * p
    a = estimation_matrix[i]
    su = 0
    for j in range(l, r):
        su += a[j-50]
        if su > p:
            return j / 100
    return MI[i] + (MA[i] - MI[i]) * p
def expected(i):
    l, r = int(MI[i] * 100 + 0.999), int(MA[i] * 100) + 1
    if MA[i] - MI[i] < 0.01 or l >= r - 1:
        return MI[i] + (MA[i] - MI[i]) * 0.5
    a = estimation_matrix[i]
    su = 0
    ret = 0
    for j in range(l, r):
        su += a[j-50]
        ret += a[j-50] * (j / 100)
    if 0.99 < su < 1.01:
        return max(MA[i], min(MI[i], ret))
    return MI[i] + (MA[i] - MI[i]) * 0.5
def receive():
    global P
    global R
    global money
    global MI, MA
    global sell_count, estimation_matrix
    if DEBUG:
        sell_count = [min(R[i], int(R[i] ** 0.5 * 1.05 ** P[i] * D[i] * (0.75 + random() * 0.5))) for i in range(N)]
        money += 1000 * sum(sell_count)
    else:
        money = int(input())
        sell_count = [int(a) for a in input().split()]
    
    for i in range(N):
        if R[i]:
            cn = R[i] ** 0.5 * 1.05 ** P[i]
            ma = (sell_count[i] + 1) / (cn * 0.75)
            mi = sell_count[i] / (cn * 1.25)
            MI[i] = max(mi, MI[i])
            if sell_count[i] < R[i]:
                MA[i] = min(ma, MA[i])
            
            l, r = int(MI[i] * 100 + 0.999), int(MA[i] * 100) + 1
            sum_prob = 0
            for j in range(l, r):
                low = max(0.75, sell_count[i] / (cn * (j / 100)))
                if sell_count[i] < R[i]:
                    high = min(1.25, (sell_count[i] + 1) / (cn * (j / 100)))
                else:
                    high = 1.25
                prob = max(0, high - low)
                estimation_matrix[i][j-50] *= prob
                sum_prob += estimation_matrix[i][j-50]
            if sum_prob:
                for j in range(l, r):
                    estimation_matrix[i][j-50] /= sum_prob
            else:
                if 0 and DEBUG:
                    assert l >= r - 2, (t, i, l, r)
            
    for i in range(N):
        if R[i]:
            if sell_count[i] * 10 >= 3 * R[i]:
                P[i] = min(P[i] + 1, 60)
            elif sell_count[i] * 10 < R[i]:
                P[i] = max(P[i] - 1, -60)
    
    if DEBUG:
        for i in range(N):
            R[i] -= sell_count[i]
    else:
        for i, a in enumerate(input().split()):
            P[i] = int(a)
        for i, a in enumerate(input().split()):
            R[i] = int(a)
    
    for i in range(N):
        total_sell_count[i] += sell_count[i]
def adj(n):
    if n < 4:
        return 3
    if n % 10 < 3:
        return n // 10 * 10
    if n % 10 < 6:
        return n // 10 * 10 + 3
    return n // 10 * 10 + 6
    
def Order(t0, f1, f2, f3, f4, f5, f6, f7, g1, g2):
    global P
    global R
    global money
    global best_allocation_flg
    target = 10
    buy_count = [0] * N
    exp = [expected(i) for i in range(N)]
    best_allocation = [(exp[i] * 1.05 ** P[i]) ** 2 for i in range(N)]
    su = sum(best_allocation)
    for i in range(N):
        best_allocation[i] *= 10 ** 8
        best_allocation[i] = adj(int(best_allocation[i]))
    best_allocation_flg = [0] * N
    for i in range(N):
        # a30 = (10 / 3 * 1.05 ** P[i] * (MA[i] * 0.0 + MI[i] * 1.0)) ** 2
        # b30 = (10 / 3 * 1.05 ** P[i] * (MA[i] * 0.5 + MI[i] * 0.5)) ** 2
        # c30 = (10 / 3 * 1.05 ** P[i] * (MA[i] * 0.1 + MI[i] * 0.9)) ** 2
        if t == 0:
            target = t0 * 10 // 3
        elif t < 5:
            target = adj(int((10 / 3 * 1.05 ** P[i] * estimate(i, 0.1) * f1) ** 2))
        elif t < 12:
            target = adj(int((10 / 3 * 1.05 ** P[i] * estimate(i, 0.1) * f2) ** 2))
        elif t < 20:
            target = adj(int((10 / 3 * 1.05 ** P[i] * exp[i] * f3) ** 2))
        elif t < (20 * 0.75 + g2 * 0.25) and P[i] < g1:
            target = adj(int((10 / 3 * 1.05 ** P[i] * exp[i] * f4) ** 2))
        elif t < (20 * 0.5 + g2 * 0.5) and P[i] < g1:
            target = adj(int((10 / 3 * 1.05 ** P[i] * exp[i] * f5) ** 2))
        elif t < (20 * 0.25 + g2 * 0.75) and P[i] < g1:
            target = adj(int((10 / 3 * 1.05 ** P[i] * exp[i] * f6) ** 2))
        elif t < g2 and P[i] < g1:
            target = adj(int((10 / 3 * 1.05 ** P[i] * exp[i] * f7) ** 2))
        elif t < 100:
            target = best_allocation[i]
            best_allocation_flg[i] = 1
        else:
            assert 0
        buy_count[i] = max(0, target - R[i])
    
    if 1:
        su_best = sum([a for a, f in zip(best_allocation, best_allocation_flg) if f])
        su_R = sum([r for r, f in zip(R, best_allocation_flg) if f])
        used = sum([max(t - r, 0) for t, r, f in zip(buy_count, R, best_allocation_flg) if not f])
        avail = max(0, money // 500 - used + su_R)
        for i in range(N):
            if best_allocation_flg[i]:
                buy_count[i] = adj(int(best_allocation[i] / su_best * avail - R[i]))
                # buy_count[i] = adj(int(best_allocation[i] / su_best * avail))
    
    while 1:
        r = sum(buy_count) * 500 / money
        if r <= 1:
            break
        
        for i in range(N):
            buy_count[i] = adj(int(buy_count[i] / r))
    
    money -= sum(buy_count) * 500
    if money < 0:
        while 1:
            pass
    assert money >= 0
    for i in range(N):
        R[i] += buy_count[i]
    if not DEBUG:
        print(1, *buy_count)
    return buy_count
def Advertise(x):
    global money
    money -= 500000 << x - 1
    if money < 0:
        while 1:
            pass
    assert money >= 0
    for i in range(N):
        P[i] = min(P[i] + x, 60)
    if not DEBUG:
        print(2, x)
def round3(x):
    return round(x * 1000) / 1000
def main_loop(t0, adv_l2, adv_r2, adv_m2_0, adv_m2_25, adv_m2_50, adv_l3, adv_r3, adv_m3, f1, f2, f3, f4, f5, f6, f7, g1, g2):
    global T, N, money
    global MI, MA, P, R, t, D, total_sell_count, LOOP
    global total_total
    ret = 0
    main_loop_loop = 1000
    for _ in range(main_loop_loop):
        ret += main(t0, adv_l2, adv_r2, adv_m2_0, adv_m2_25, adv_m2_50, adv_l3, adv_r3, adv_m3, f1, f2, f3, f4, f5, f6, f7, g1, g2)
    return ret / main_loop_loop
def main(t0, adv_l2, adv_r2, adv_m2_0, adv_m2_25, adv_m2_50, adv_l3, adv_r3, adv_m3, f1, f2, f3, f4, f5, f6, f7, g1, g2):
    global T, N, money
    global MI, MA, P, R, t, D, total_sell_count, LOOP
    global total_total
    global sell_count, best_allocation_flg
    if DEBUG:
        T, N, money = 52, 10, 2 * 10 ** 6
        D = [random() + 0.5 for _ in range(N)]
    else:
        T, N, money = map(int, input().split())
        D = [random() + 0.5 for _ in range(N)]
    total_sell_count = [0] * N
    sell_count = [0] * N
    best_allocation_flg = [0] * N
    MA = [1.5] * N
    MI = [0.5] * N
    P = [0] * N
    R = [0] * N
    max_buy_count = 30
    for t in range(T):
        adv_m2 = adv_m2_0 * (1 - t / 25) + adv_m2_25 * (t / 25) if t <= 25 else adv_m2_25 * (1 - (t - 25) / 25) + adv_m2_50 * ((t - 25) / 25)
        if (adv_l3 < t < adv_r3 and money >= adv_m3) and sorted(P)[-1] < 58:
            Advertise(3)
        elif 1 and (adv_l2 < t < adv_r2 and money >= adv_m2) and sorted(P)[-1] < 59:
            Advertise(2)
        elif 0 and (adv_l2 < t < adv_r2 and money - 1000000 >= max_buy_count * 1.0 * 500) and sorted(P)[-1] < 59:
            Advertise(2)
        elif (0 < t < 0 and money >= 1200000) and sorted(P)[-1] < 59:
            Advertise(2)
        elif t < 4 and money >= 6000000:
            Advertise(4)
        elif t < 4 and money >= 3000000:
            Advertise(3)
        else:
            buy_count = Order(t0, f1, f2, f3, f4, f5, f6, f7, g1, g2)
            max_buy_count = max(max_buy_count, sum(buy_count))
        if DEBUG:
            if LOOP == 1:
                print("-" * 20)
                print("t =", t)
                print("Money =", money)
                
        receive()
        if DEBUG:
            if LOOP == 1:
                print("Money =", money)
                print("P =", P)
                print("R =", R)
                print("sell_count =", sell_count)
                print("MI =", ["{:.2f}".format(a) for a in MI])
                print("MA =", ["{:.2f}".format(a) for a in MA])
                print("D  =", ["{:.2f}".format(a) for a in D])
                print("Flg =", best_allocation_flg)
                print("Total =", sum(total_sell_count), total_sell_count)
                
                for i in range(10):
                    l, r = int(MI[i] * 100 + 0.999), int(MA[i] * 100) + 1
                    L = [estimation_matrix[i][j-50] for j in range(l, r)]
                    print("i =", i)
                    print(round3(MI[i]), round3(MA[i]), round3(D[i]), l / 100, (r - 1) / 100, [round3(a) for a in L], round3(sum(L)))
                    print(estimate(i, 0.1), estimate(i, 0.5))
                
    if LOCAL:
        total_total += sum(total_sell_count)
    if LOCAL and 4 >= LOOP > 1:
        print("loop =", loop, total_sell_count, sum(total_sell_count), round(total_total / (loop + 1)))
    return sum(total_sell_count)
def objective(trial):
    t0 = trial.suggest_int('t0', 1, 1)
    adv_l2 = trial.suggest_int('adv_l2', 1, 1)
    adv_r2 = trial.suggest_int('adv_r2', 33, 33)
    # adv_m2 = trial.suggest_float('adv_m2', 105 * 10 ** 4, 130 * 10 ** 4)
    adv_m2_0 = trial.suggest_float('adv_m2_0', 105 * 10 ** 4, 150 * 10 ** 4)
    adv_m2_25 = trial.suggest_float('adv_m2_25', 105 * 10 ** 4, 150 * 10 ** 4)
    adv_m2_50 = trial.suggest_float('adv_m2_50', 105 * 10 ** 4, 150 * 10 ** 4)
    adv_l3 = trial.suggest_int('adv_l3', 16, 16)
    adv_r3 = trial.suggest_int('adv_r3', 50, 50)
    adv_m3 = trial.suggest_float('adv_m3', 220 * 10 ** 4, 300 * 10 ** 4)
    f1 = trial.suggest_float('f1', 0.70, 0.8)
    f2 = trial.suggest_float('f2', 0.60, 0.9)
    f3 = trial.suggest_float('f3', 0.50, 0.9)
    f4 = trial.suggest_float('f4', 0.50, 0.9)
    f5 = trial.suggest_float('f5', 0.50, 1.0)
    f6 = trial.suggest_float('f6', 0.50, 1.5)
    f7 = trial.suggest_float('f7', 0.80, 2.0)
    g1 = trial.suggest_int('g1', 40, 60)
    g2 = trial.suggest_int('g2', 40, 50)
    score = main_loop(t0, adv_l2, adv_r2, adv_m2_0, adv_m2_25, adv_m2_50, adv_l3, adv_r3, adv_m3, f1, f2, f3, f4, f5, f6, f7, g1, g2)
    return -score
N = 10
sell_count = [0] * N
best_allocation_flg = [0] * N
estimation_matrix = [[1 / 101] * 101 for _ in range(N)]
if OP:
    T, N, money = None, None, None
    MI, MA, D = None, None, None
    P, R, t, total_sell_count = None, None, None, None
    study = optuna.create_study()
    study.optimize(objective, n_trials=1000)
    print("Best Params =", study.best_params)
    print("Best Score =", study.best_value)
else:
    total_total = 0
    _t = 0
    Params = {
        'adv_l2': 1, 'adv_r2': 33, 'adv_m2': 1150894.1177157715, 
        'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2946328.352258085, 
        'f1': 0.7577663544011545, 'f2': 0.4828227141584538, 'f3': 0.4972439506386438, 
        'f4': 0.5629233324271415, 'f5': 0.5131416541498899, 'f6': 0.45233343537815346, 
        'f7': 0.438522644272038, 'g1': 45, 'g2': 44}
    Params = {
        'adv_l2': 1, 'adv_r2': 33, 
        'adv_m2_0': 1123020.3753850544, 'adv_m2_25': 1288743.5426866524, 'adv_m2_50': 1253489.0596099684, 
        'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2844417.54023583, 
        'f1': 0.7205992286224824, 'f2': 0.5565542576756614, 'f3': 0.4399660197983403, 
        'f4': 0.5212192343140634, 'f5': 0.5250031287694346, 'f6': 0.23877967298253283, 
        'f7': 0.10991904486471206, 'g1': 44, 'g2': 48}
    Params = {'t0': 1, 'adv_l2': 1, 'adv_r2': 33, 'adv_m2_0': 1426078.8881682719, 'adv_m2_25': 1283126.1323284279, 'adv_m2_50': 1294982.9461560582, 
        'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2793228.676762244, 'f1': 0.9253559437203199, 'f2': 1.0782986213593686, 'f3': 0.8903959239285214, 'f4': 
        1.0546976032171025, 'f5': 1.0111526161355209, 'f6': 0.710258091945632, 'f7': 0.7596096884833397, 'g1': 40, 'g2': 45}
    Params = {
        't0': 1, 'adv_l2': 1, 'adv_r2': 33, 
        'adv_m2_0': 1110853.9703218008, 'adv_m2_25': 1265411.1806625545, 'adv_m2_50': 1174453.8497893137, 
        'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2835338.226649481, 
        'f1': 0.7959737776095417, 'f2': 0.6711256267655628, 'f3': 0.5458483694117402, 
        'f4': 0.6985138326712955, 'f5': 0.8177103445792984, 'f6': 1.1195154172991402, 
        'f7': 1.063274260974727, 'g1': 48, 'g2': 48}
    Params = {'t0': 1, 'adv_l2': 1, 'adv_r2': 33, 
              'adv_m2_0': 1085794.3113618647, 'adv_m2_25': 1364320.0322415212, 'adv_m2_50': 1264407.0072151911, 
              'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2689021.087630408, 
              'f1': 0.7591365852771064, 'f2': 0.6568779372367627, 'f3': 0.5365137179442521, 
              'f4': 0.7301887169857361, 'f5': 0.729854029317634, 'f6': 0.6391223277425482, 
              'f7': 1.4121129251060212, 'g1': 44, 'g2': 42}
    Params = {'t0': 1, 'adv_l2': 1, 'adv_r2': 33, 'adv_m2_0': 1070701.297158546, 'adv_m2_25': 1050631.533440852, 'adv_m2_50': 1309850.6325924492, 
        'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2987246.773087814, 'f1': 0.7178026016906326, 'f2': 0.6914188178555515, 'f3': 0.6086726301173585, 'f4': 
        0.6431800953698513, 'f5': 0.7405037759739107, 'f6': 0.6778052075743504, 'f7': 1.7592543252021107, 'g1': 45, 'g2': 50}
    for k, v in Params.items():
        exec(k + str(" = ") + str(v))
    t0 = 1
    for loop in range(LOOP):
        _t += main(t0, adv_l2, adv_r2, adv_m2_0, adv_m2_25, adv_m2_50, adv_l3, adv_r3, adv_m3, f1, f2, f3, f4, f5, f6, f7, g1, g2)
    if LOCAL:
        print("LOOP =", LOOP, round(total_total / LOOP), round(_t / LOOP))
_ = '''
[I 2023-10-02 16:11:43,324] Trial 228 finished with value: -185364.995 and parameters: {'t0': 1, 'adv_l2': 1, 'adv_r2': 33, 'adv_m2_0': 1085794
    .3113618647, 'adv_m2_25': 1364320.0322415212, 'adv_m2_50': 1264407.0072151911, 'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2689021.087630408, 'f1': 0
    .7591365852771064, 'f2': 0.6568779372367627, 'f3': 0.5365137179442521, 'f4': 0.7301887169857361, 'f5': 0.729854029317634, 'f6': 0
    .6391223277425482, 'f7': 1.4121129251060212, 'g1': 44, 'g2': 42}. Best is trial 228 with value: -185364.995.
Best Params = {'t0': 1, 'adv_l2': 1, 'adv_r2': 33, 'adv_m2_0': 1110853.9703218008, 'adv_m2_25': 1265411.1806625545, 'adv_m2_50': 1174453.8497893137, 
    'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2835338.226649481, 'f1': 0.7959737776095417, 'f2': 0.6711256267655628, 'f3': 0.5458483694117402, 'f4': 0
    .6985138326712955, 'f5': 0.8177103445792984, 'f6': 1.1195154172991402, 'f7': 1.063274260974727, 'g1': 48, 'g2': 48}
Best Score = -185012.96
Best Params = {'t0': 1, 'adv_l2': 1, 'adv_r2': 33, 'adv_m2_0': 1426078.8881682719, 'adv_m2_25': 1283126.1323284279, 'adv_m2_50': 1294982.9461560582, 
    'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2793228.676762244, 'f1': 0.9253559437203199, 'f2': 1.0782986213593686, 'f3': 0.8903959239285214, 'f4': 1
    .0546976032171025, 'f5': 1.0111526161355209, 'f6': 0.710258091945632, 'f7': 0.7596096884833397, 'g1': 40, 'g2': 45}
Best Score = -61533.66
Best Params = {'adv_l2': 1, 'adv_r2': 33, 'adv_m2_0': 1123020.3753850544, 'adv_m2_25': 1288743.5426866524, 'adv_m2_50': 1253489.0596099684, 'adv_l3': 
    16, 'adv_r3': 50, 'adv_m3': 2844417.54023583, 'f1': 0.7205992286224824, 'f2': 0.5565542576756614, 'f3': 0.4399660197983403, 'f4': 0
    .5212192343140634, 'f5': 0.5250031287694346, 'f6': 0.23877967298253283, 'f7': 0.10991904486471206, 'g1': 44, 'g2': 48}
Best Score = -183440.27
Best Params = {'adv_l2': 1, 'adv_r2': 33, 'adv_m2': 1150894.1177157715, 'adv_l3': 16, 'adv_r3': 50, 'adv_m3': 2946328.352258085, 'f1': 0
    .7577663544011545, 'f2': 0.4828227141584538, 'f3': 0.4972439506386438, 'f4': 0.5629233324271415, 'f5': 0.5131416541498899, 'f6': 0
    .45233343537815346, 'f7': 0.438522644272038, 'g1': 45, 'g2': 44}
Best Score = -182240.325
Best Params = {
'adv_l2': 1, 'adv_r2': 33, 'adv_m2': 1059258.2189775354, 
'adv_l3': 18, 'adv_r3': 48, 'adv_m3': 2686445.781066581, 
'f1': 0.7977800220202227, 'f2': 0.551032831198818, 'f3': 0.5299220276322835, 
'f4': 0.5498393778548679, 'f5': 0.42812314931333745, 'f6': 0.18079451958248885, 
'f7': 0.1333551668112509, 'g1': 41, 'g2': 47}
Best Score = -179423.575
Best Params = {
'adv_l2': 3, 'adv_r2': 34, 'adv_m2': 1077583.6894016732, 
'adv_l3': 15, 'adv_r3': 39, 'adv_m3': 2641221.6852187263, 
'f1': 0.6608625427038423, 'f2': 0.536255559402984, 'f3': 0.544224068328003, 
'f4': 0.5930760450080326, 'f5': 0.5296353385134165, 'f6': 0.4997011249487702, 
'f7': 0.18112500260814665, 'g1': 51, 'g2': 50}
Best Score = -179557.162
Best Params = {
'adv_l2': 4, 'adv_r2': 34, 'adv_m2': 1186197.4406347682, 
'adv_l3': 18, 'adv_r3': 43, 'adv_m3': 2931408.6158438902, 
'f1': 0.8329027232431818, 'f2': 0.4383836146967173, 'f3': 0.5335765817506617, 
'f4': 0.5582777882317832, 'f5': 0.48926189592932584, 'f6': 0.3277355570137359, 
'f7': 0.19762882186386968, 'g1': 48, 'g2': 44}
Best Params = {
'adv_l2': 1, 'adv_r2': 34, 'adv_m2': 1071996.048402824, 
'adv_l3': 18, 'adv_r3': 39, 'adv_m3': 2506057.429736735, 
'f1': 0.7530461775247944, 'f2': 0.5784067161547253, 'f3': 0.3998962794048304, 
'f4': 0.5294730042963244, 'f5': 0.4904776465015429, 'f6': 0.3933636958636354, 
'f7': 0.2161664699683005, 'g1': 50}
Best Score = -174102.633
Best Params = {'adv_l3': 12.316204575888603, 'adv_r3': 47.71022165578954, 
'adv_m': 2942277.919106235, 'f1': 0.7837350220521555, 'f2': 0.5465937999426904, 
'f3': 0.46522777277350125, 'f4': 0.4715163081906667, 'f5': 0.5818274160156429, 
'f6': 0.1329005755644884, 'f7': 0.453691378873825, 'g1': 44}
Best Score = -176159.62
'''
הההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההההה
XXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXXX
0