// Policy iteration bug fixed in revision 7707

mdp

module strips
state : [ 0 .. 4 ] init 0;

[ takestick ] state=0 -> 1: (state'=1) ;
[ askforhelp ] state=0 -> 0.4: (state'=2)  + 0.6: (state'=0) ;
[ takestick ] state=2 -> 1: (state'=3) ;
[ leavehelp ] state=2 -> 1: (state'=0) ;
[ takebananas2 ] state=3 -> 0.2: (state'=4)  + 0.8: (state'=3) ;
[ dropstick ] state=3 -> 1: (state'=2) ;
[ leavehelp ] state=3 -> 1: (state'=1) ;
[ askforhelpwithstick ] state=1 -> 0.8: (state'=3)  + 0.2: (state'=1) ;
[ askforhelp ] state=1 -> 0.4: (state'=3)  + 0.6: (state'=1) ;
[ dropstick ] state=1 -> 1: (state'=0) ;
endmodule

rewards "cost"
[ takestick ] true : 5;
[ askforhelp ] true : 8;
[ leavehelp ] true : 1;
[ takebananas2 ] true : 3;
[ askforhelpwithstick ] true : 8;
[ dropstick ] true : 1;
endrewards