<P>
<br></P><FONT face=dcr10 size=1>
<P align=left><FONT size=3>In repeated games there is in general a large set of equilibria. We also know that</FONT></P>
<P align=left><FONT size=3>in the repeated prisoners dilemma there is a profusion of neutrally stable strategies,</FONT></P>
<P align=left><FONT size=3>but no strategy that is evolutionarily stable. This paper investigates whether and</FONT></P>
<P align=left><FONT size=3>how neutrally stable strategies can be upset in a process of mutation and selection.</FONT></P>
<P align=left><FONT size=3>While neutral stability excludes that mutants have a selective advantage themselves, it</FONT></P>
<P align=left><FONT size=3>does not rule out the possibility that mutants that are neutral can enter a population</FONT></P>
<P align=left><FONT size=3>and create a selective advantage for a second mutant. This will be called an indirect</FONT></P>
<P align=left><FONT size=3>invasion and the central results show that, for high enough continuation probability,</FONT></P>
<P align=left><FONT size=3>there is no strategy that is robust against indirect invasions. Such stepping stone paths</FONT></P>
<P align=left><FONT size=3>out of equilibrium generally exist both in the direction of more and in the direction of</FONT></P>
<P><FONT size=3>less cooperation.</FONT></P></FONT>
[此贴子已经被作者于2007-3-2 2:55:28编辑过]