Attacks on Security Protocols

Consider the following $4-$message protocol:

$1.$ A$->$S: $($B$,\{($A$,$ K$1)\}$KpbS$)$

$2.$ S$->$B: A

$3.$ B$->$S: $($A$,\{($B$,$ K$2)\}$KpbS$)$

$4.$ S$->$A: $($B$,\{$K$2\}$K$1)$

And the following attack trace:

$1.$ I$($A$)->$S: $($B$,\{($A$,$ K$)\}$KpbS$)$

$2.$ S$->$ B: A

$3.$ B$->$S: $($A$,\{($B$,$ K$2)\}$KpbS$)$

$4.$ S$->$I$($A$)$: $($B$,\{$K$2\}$K$)$

Which one of the following changes to the messages of the protocol specification would fix the attack above $($i$.$e$.$ such that the attack trace then becomes impossible$)$ while also maintain a correct protocol that would still achieve the purpose that the original protocol specification above achieves:

a$)3.$ B$->$S: $($A$,\{($B$,\{$K$2\}$KpbA$)\}$KpbS$)$

b$)4.$ S$->$A: $($B$,\{$K$2,$ A$\}$K$1)$

c$)2.$ S$->$B: $\{$A$\}$KpbB

d$)2.$ S$->$B: B

e$)3.$ B$->$S: $($A$,\{($B$,\{$K$2\}$KprS$)\}$KpbS$)$

f$)1.$ A$->$S: $\{($B$,$ A$,$ K$1)\}$KpbS

g$)1.$ A$->$S: $($A$,\{($B$,$ K$1)\}$KpbS$)$

h$)4.$ S$->$A: $($B$,\{$K$1\}$K$2)$

i$)4.$ S$->$A: $($A$,$ B$,\{$K$2\}$K$1)$

j$)2.$ S$->$B: A$,$ B