Update for C++11
As of the C++11 standard, string-to-number conversion and vice-versa are built in into the standard library. All the following functions are present in <string> (as per paragraph 21.5).
string to numeric
float stof(const string& str, size_t *idx = 0);
double stod(const string& str, size_t *idx = 0);
long double stold(const string& str, size_t *idx = 0);
int stoi(const string& str, size_t *idx = 0, int base = 10);
long stol(const string& str, size_t *idx = 0, int base = 10);
unsigned long stoul(const string& str, size_t *idx = 0, int base = 10);
long long stoll(const string& str, size_t *idx = 0, int base = 10);
unsigned long long stoull(const string& str, size_t *idx = 0, int base = 10);
Each of these take a string as input and will try to convert it to a number. If no valid number could be constructed, for example because there is no numeric data or the number is out-of-range for the type, an exception is thrown (std::invalid_argument or std::out_of_range).
If conversion succeeded and idx is not 0, idx will contain the index of the first character that was not used for decoding. This could be an index behind the last character.
Finally, the integral types allow to specify a base, for digits larger than 9, the alphabet is assumed (a=10 until z=35). You can find more information about the exact formatting that can parsed here for floating-point numbers, signed integers and unsigned integers.
Finally, for each function there is also an overload that accepts a std::wstring as it's first parameter.
numeric to string
string to_string(int val);
string to_string(unsigned val);
string to_string(long val);
string to_string(unsigned long val);
string to_string(long long val);
string to_string(unsigned long long val);
string to_string(float val);
string to_string(double val);
string to_string(long double val);
These are more straightforward, you pass the appropriate numeric type and you get a string back. For formatting options you should go back to the C++03 stringsream option and use stream manipulators, as explained in an other answer here.
As noted in the comments these functions fall back to a default mantissa precision that is likely not the maximum precision. If more precision is required for your application it's also best to go back to other string formatting procedures.
There are also similar functions defined that are named to_wstring, these will return a std::wstring.
