>1. What is the reward assignment for N_h and M_h in case 2 (slides 203) > >For N_h, can I use mark(SH) + mark(H)? >For M_h, can I use if(mark(RH) == 0 && mark(SL) < alpha * alpha) return 1; >else return 0; N_h is the departure rate of high priority clients. Mark("SH")/alpha +mark("H") will only give you the # of high priority clients in the system, which is not correct. The correct answer is: (mark("SH")/alpha) * mu + mark("H") * mu, or simply rate("T3") + rate("T9"). M_h again is the rejection rate, not the rejection probability. So you should return lambda_h, not just 1. Also your conditions are not right. It should be: if (mark("RH")==0 && mark("SH")==0) && mark("SL") < alpha * alpha) return lambda_h; else return 0; >2. Any other way to calculate MTTA except pr_mtta in slide 180? Yes with the reward assignment defined by reliability() on slide #180, you can also calculate the cumulative reward till absorption by: pr_cum_abs("mean time to absorption = ", reliability); >For 1 above, one thing still confused is the reward for M_h. >Why is mark("SH")==0, not mark("RS") < alpha? If that, the reward is: >if (mark("RH")==0 && mark("RS") < alpha) && mark("SL) < alpha * alpha) >return lambda_h; else return 0; mark("RS") < alpha will have the same effect as mark("RS")==0, because if there are less than alpha tokens in place RS, the immediate transitions (for QoS upgrade of low-priority clients) will consume all such less than alpha tokens immediately. So in effect mark("RS") will just contain 0 tokens. > Another question for paper 3 Effect of Intrusion Detection and Response > on Reliability of Cyber Physical Systems. > > Is the reliability reward as below? > > If(mark("energy") == 0 || mark("N_g") <= 2 * mark("N_b") || > mark("impaired") >= 1) return 1; > > Else return 0; > > > > Is the MTTF as below? > > pr_cum_abs("mean time to absorption = ", reliability); > > > > Is there any method like pr_mtta() to calculate the MTTF directly? Yes your reliability reward function is correct! You can also use the built-in function pr_mtta() to calculate the MTTF since pr_mtta() will return the average time to absorption of this SPN model, which happens when any of three failure conditions is true.