// Running example MDP (see Figure 2, p.6) mdp module M // s=i for state s_i s : [0..3]; [go] s=0 -> (s'=1); [safe] s=1 -> 0.7:(s'=0) + 0.3:(s'=2); [risk] s=1 -> 0.5:(s'=2) + 0.5:(s'=3); [finish] s=2 -> (s'=2); [stop] s=3 -> (s'=3); [reset] s=3 -> (s'=0); endmodule rewards // state rewards s=0 : 1; s=1 : 2; // transition rewards [go] s=0 : 1; [safe] s=1 : 1; [risk] s=1 : 4; [reset] s=3 : 5; endrewards label "initial" = s=0; label "succ" = s=2; label "fail" = s=3;