Asymmetric algorithms use pairs of keys. One is used for encryption and the other one for decryption. The decryption key is typically kept secretly, therefore called ``private key'' or ``secret key'', while the encryption key is spread to all who might want to send encrypted messages, therefore called ``public key''. Everybody having the public key is able to send encrypted messages to the owner of the secret key. The secret key can't be reconstructed from the public key. The idea of asymmetric algorithms was first published 1976 by Diffie and Hellmann.
Asymmetric algorithms seem to be ideally suited for real-world use: As the secret key does not have to be shared, the risk of getting known is much smaller. Every user only needs to keep one secret key in secrecy and a collection of public keys, that only need to be protected against being changed. With symmetric keys, every pair of users would need to have an own shared secret key. Well-known asymmetric algorithms are RSA, DSA, ELGAMAL.
However, asymmetric algorithms are much slower than symmetric ones. Therefore, in many applications, a combination of both is being used. The asymmetric keys are used for authentication and after this has been successfully done, one or more symmetric keys are generated and exchanged using the asymmetric encryption. This way the advantages of both algorithms can be used. Typical examples of this procedure are the RSA/IDEA combination of PGP2 or the DSA/BLOWFISH used by GnuPG.