UVa 10057 A mid-summer night's dream. (二分&可以取哪些作为中位数?)

本文探讨了在2200年科学进步的背景下,人们如何通过直接连接人脑与计算机CPU实现观看他人梦境的技术。面对依赖计算机导致的分析能力退化问题,科学家在中暑之夜梦到一系列无符号整数,并运用这些数字来寻找组合锁的最佳代码。文章详细阐述了解决此类问题的算法和输入输出规范。

10057 - A mid-summer night's dream.

Time limit: 30.000 seconds 

http://uva.onlinejudge.org/index.php?option=com_onlinejudge&Itemid=8&category=113&page=show_problem&problem=998

This is year 2200AD. Science has progressed a lot in two hundred years. Two hundred years is mentioned here because this problem is being sent back to 2000AD with the help of time machine. Now it is possible to establish direct connection between man and computer CPU. People can watch other people’s dream on 3D displayer (That is the monitor today) as if they were watching a movie. One problem in this century is that people have become so dependent on computers that their analytical ability is approaching zero. Computers can now read problems and solve them automatically. But they can solve only difficult problems. There are no easy problems now. Our chief scientist is in great trouble as he has forgotten the number of his combination lock. For security reasons computers today cannot solve combination lock related problems. In a mid-summer night the scientist has a dream where he sees a lot of unsigned integer numbers flying around. He records them with the help of his computer, Then he has a clue that if the numbers are (X1, X2,   …  , Xn) he will have to find an integer number A (This A is the combination lock code) such that

             

             (|X1-A| + |X2-A| + … … + |Xn-A|) is minimum.

 

Input

Input will contain several blocks. Each block will start with a number n (0<n<=1000000) indicating how many numbers he saw in the dream. Next there will be n numbers. All the numbers will be less than 65536. The input will be terminated by end of file.

 

Output

For each set of input there will be one line of output. That line will contain the minimum possible value for A. Next it will contain how many numbers are there in the input that satisfy the property of A (The summation of absolute deviation from A is minimum). And finally you have to print how many possible different integer values are there for A (these values need not be present in the input). These numbers will be separated by single space.

 

Sample Input:

2
10
10
4
1
2
2
4

Sample Output:

10 2 1
2 2 1


输出注意:

第一个数是最小的A,第二个数是在数列{Xn}中有多少个数可以是A,第三个数是在有多少个数可以是A。


完整代码:

/*0.245s*/

#include<cstdio>
#include<algorithm>
using namespace std;
#define sf scanf
#define pf printf

int a[1000005];

int main()
{
	int n, m, i;
	while (~sf("%d", &n))
	{
		m = (n - 1) >> 1;
		for (i = 0; i < n; ++i)
			sf("%d", &a[i]);
		sort(a, a + n);
		pf("%d %d %d\n", a[m], ((n & 1) == 0 ? upper_bound(a, a + n, a[m + 1]) : upper_bound(a, a + n, a[m])) - lower_bound(a, a + n, a[m]), ((n & 1) == 0 && a[m] != a[m + 1] ? a[m + 1] - a[m] + 1 : 1));
	}
	return 0;
}


### GIB-UVA ERP-BCI HDF5 文件格式及其处理方法 HDF5 是一种用于存储大量科学数据的文件格式,广泛应用于神经科学研究领域。对于 GIB-UVA ERP-BCI 数据集中的 HDF5 文件,通常包含了脑电图(EEG)信号以及其他元数据信息。以下是关于该类文件的一些重要细节以及如何对其进行处理的方法。 #### 1. HDF5 文件结构概述 HDF5 文件是一种分层的数据存储格式,类似于文件系统的目录树结构。它支持多种数据类型,包括数组、表格和字符串等。在 GIB-UVA ERP-BCI 的上下文中,这些文件可能包含以下内容: - **实验记录**:如时间戳、采样率和其他实验参数。 - **原始 EEG 数据**:多通道的时间序列数据。 - **事件标记**:表示刺激呈现或其他行为事件的时间点。 这种层次化的结构使得研究人员可以轻松访问特定部分的数据而无需加载整个文件[^3]。 #### 2. 处理 HDF5 文件所需的工具 为了读和操作 HDF5 文件,可以使用 Python 中的 `h5py` 或 MATLAB 提供的相关库。下面是一个简单的例子展示如何利用 `h5py` 打开并探索一个 HDF5 文件的内容: ```python import h5py def explore_hdf5(file_path): with h5py.File(file_path, &#39;r&#39;) as f: print(&quot;Keys:&quot;, list(f.keys())) # 列出顶层组名 for key in f.keys(): item = f[key] if isinstance(item, h5py.Dataset): print(f&quot;{key} is a dataset with shape {item.shape}&quot;) elif isinstance(item, h5py.Group): print(f&quot;{key} is a group containing:&quot;) for sub_key in item.keys(): print(f&quot; - {sub_key}&quot;) explore_hdf5(&#39;example.h5&#39;) ``` 上述脚本会打印出给定 HDF5 文件的所有顶级键,并区分它们是数据集还是子组[^4]。 #### 3. 内存管理注意事项 如果尝试运行某些大型模型(例如 DeepSeek-R1),可能会遇到内存不足的情况,正如引用中提到的例子所示[^2]。在这种情况下,建议采以下措施来优化资源分配: - 使用更高效的算法减少计算需求; - 增加物理 RAM 或启用虚拟内存扩展; - 对于 GPU 加速环境,考虑调整批次大小或切换到较低精度浮点数运算模式(FP16 vs FP32)。 此外,在处理大尺寸的 HDF5 文件时也需要注意类似的性能瓶颈问题&mdash;&mdash;可以通过逐块加载而非一次性全部载入的方式来缓解这一挑战[^5]。 #### 4. 特殊情况下的预处理技术 针对 BCI 应用场景下采集得到的高维时空域特征矩阵,往往还需要执行一系列标准化流程,比如去噪滤波器应用、基线校正以及重参考变换等等。具体实现决于实际研究目标和个人偏好设置等因素影响。 --- ###
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值