Path-Wise Test Data Generation Based on Heuristic Look-Ahead Methods

Algorithm 1

Best-first-search branch and bound.

Input   𝑝: the path to be traversed

Output  result ⟨𝑉𝑎𝑟𝑖𝑎𝑏𝑙𝑒,𝑉𝑎𝑙𝑢𝑒⟩: the test data making 𝑝 feasible

Stage 1: Initialization

(1)   call Algorithm Irrelevant variable removal;

(2)   result ←  null;

(3)   call Algorithm Maintaining path consistency;

(4)   call Algorithm Dynamic variable ordering;

(5)   call Algorithm Path tendency calculation;

(6)   𝑥₁ ← head (𝑄₁);

(7)   call Algorithm Initial domain calculation;

(8)   𝑉₁₁ ← select (𝐷₁₁);

(9)   initial state ← (null, 𝑥₁,𝐷₁₁,𝑉₁₁, active, 𝑄₁);

(10)  𝑆_cur ← initial state;

Stage 2: State space search

Begin

(11)  for (Pre, 𝑥_𝑖,𝐷_𝑖𝑗,𝑉_𝑖𝑗, active, 𝑄_𝑖) (𝑖→1: 𝑛)

(12)   path consistent ←  false;

(13)   call Algorithm Maintaining path consistency;

(14)   if (path consistent = true)

(15)     𝑆_cur ← (Pre, 𝑥_𝑖,𝐷_𝑖𝑗,𝑉_𝑖𝑗, extensive, 𝑄_𝑖);

(16)     result ← result ∪𝑥_𝑖,𝑉_𝑖𝑗;

(17)     FV ← FV − {𝑥_𝑖};

(18)     PV ← PV + {𝑥_𝑖};

(19)     call Algorithm Dynamic variable ordering;

(20)     if (𝑄_𝑖 = null)

(21)       𝑆_cur ← final state;

(22)       foreach  𝑥^∗∈𝑋_irrel

(23)          result ← result ∪𝑥^∗,𝑉_random;

(24)     else  Pre ← 𝑆_cur;

(25)       𝑥_𝑖 ← head (𝑄_𝑖);

(26)       call Algorithm Initial domain calculation;

(27)       𝑉_𝑖1 ← select (𝐷_𝑖1);

(28)       𝑆_cur ← (Pre, 𝑥_𝑖,𝐷_𝑖1,𝑉_𝑖1, active, 𝑄_𝑖);

(29)   else if  𝐷_𝑖𝑗>1 && 𝑗<𝑚

(30)       call Algorithm Bisection;

(31)       𝑉_𝑖𝑗 ← select (𝐷_𝑖𝑗);

(32)       𝑆_cur ← (Pre, 𝑥_𝑖,𝐷_𝑖𝑗,𝑉_𝑖𝑗, active, 𝑄_𝑖);

(33)     else  𝑆_cur ← (Pre, 𝑥_𝑖,𝐷_𝑖𝑗,𝑉_𝑖𝑗, inactive, 𝑄_𝑖);

(34)       𝑃𝑟𝑒←𝑆_cur;

(35)       𝑆_cur ← (Pre, 𝑥_𝑖,𝐷_𝑖𝑗,𝑉_𝑖𝑗, active, 𝑄_𝑖);

(36)       PV ← PV − {𝑥_𝑖};

(37)  return  result;

End