1: \begin{abstract}

2:  The most efficient MC weights for the calculation of physical, canonical 

3:  expectation values are not necessarily those of the canonical ensemble. 

4:  The use of suitably generalized ensembles can lead to a much faster

5:  convergence of the simulation. Although not realized by nature, these 

6:  ensembles can be implemented on computers.

7:  In recent years generalized ensembles have in particular

8:  been studied for the simulation of complex systems. For these

9:  systems it is typical that conflicting constraints lead to free

10:  energy barriers, which fragment the configuration space. Examples

11:  of major interest are spin glasses and proteins. In my overview I

12:  first comment on the strengths and weaknesses of a few major

13:  approaches, multicanonical simulations, transition variable methods,

14:  and parallel tempering. Subsequently, two applications are presented:

15:  a new analysis of the Parisi overlap distribution for the 3d 

16:  Edwards-Anderson Ising

17:  spin glass and the helix-coil transition of amino-acid homo-oligomers.

18: \end{abstract}

19: