A Rough Penalty Genetic Algorithm for Multicast Routing in Mobile Ad Hoc Networks

Pseudocode 2

Pseudocode of the RP method.

conN is mean constraint numbers

popnum is population numbers

IT is information table, which is (𝑐𝑜𝑛𝑁+1)∗𝑝𝑜𝑝𝑛𝑢𝑚  matrix

gp is good population which decision variable = 1 from IT

bp is bad population which decision variable = 0 from IT

𝑔𝑐_𝑘  is good characteristic from gp

𝑏𝑐_𝑘  is bad characteristic from bp

t is the number of current generation

𝜋_𝑘(𝑡) is a RP exponent for the t generation

𝜆_𝑘 is a RP coefficient

𝐶  is constraints weight

𝜏  is next generation penalty coefficient

Begin

   Create information table 𝐼𝑇;

   //Divide 𝐼𝑇  to two group 𝑔𝑝  and 𝑏𝑝

   For i = 1 to 𝑝𝑜𝑝𝑛𝑢𝑚

    IF decision variable = 1 Pick 𝐼𝑇(𝑖)  to good population 𝑔𝑝;

    Else Pick 𝐼𝑇(𝑖)  to bad population 𝑏𝑝;

    End if

   End for

   //Find characteristic from 𝑔𝑝  and 𝑏𝑝

   For 𝑘=1 to 𝑐𝑜𝑛𝑁

    𝑔𝑐_𝑘 = mod all 𝑔𝑝_𝑘;

    𝑏𝑐_𝑘 = mod all 𝑏𝑝_𝑘;

   End for

   //Find RP coefficient 𝜇

   For 𝑘=1 to 𝑐𝑜𝑛𝑁

    If 𝑔𝑐_𝑘=𝑏𝑐_𝑘  𝜆_𝑘=𝑔𝑐_𝑘;

    Else 𝜆_𝑘=1;

    End if

   End for

   //Modify RP penalty exponent 𝜋(k,  t)

   For 𝑘=1 to 𝑐𝑜𝑛𝑁

    𝜋_𝑘(𝑡) = 𝜋_𝑘(𝑡−1)∗𝜆_𝑘;

    𝜏_𝑘=(𝐶×𝑡)∧𝜋_𝑘(𝑡), for all 𝑘=1 to 𝑐𝑜𝑛𝑁;

   End for

End