Algorithm 1

Standard Optimization Algorithm

Require:  r patient responses at each of c initial dose combinations, D={(di,yi)}i=1n0
1: nn0=r×c
2: 1STOPFALSE
3: Obtain p(fD,d~) and p(doptD) using fitted GP             ⊳ Obtain posteriors
4: d*argmind~μ(d~)+σ(d~)                       ⊳ Obtain effective best point
5: f*μ(d*)                             ⊳ Obtain best value of f
6: Calculate αAEI(d~D) for d~D                    ⊳ Compute AEI
7: while  n<N and 1STOP=FALSE  do
8:  d(c+1)argmaxd~DαAEI(d~D)                   ⊳ Obtain next dose
9:  for  i=1,,r  do
10:    Evaluate yi at d(c+1)                      ⊳ Observe outcomes
11:  end for
12:  nn+r; cc+1                       ⊳ Update n,c
13:  DD{(di(c+1),yi)}i=1r                      ⊳ Update data
14:  Obtain p(fD,d~) and p(doptD) using fitted GP           ⊳ Obtain posteriors
15:  d*argmind~μ(d~)+σ(d~)                     ⊳ Obtain effective best point
16:  f*μ(d*)                           ⊳ Obtain best value of f
17:  Calculate αAEI(d~D) for d~D                   ⊳ Compute AEI
18:  Update 1STOP using (4)                      ⊳ Update stopping rule
19: end while
Require:  r patient responses at each of c initial dose combinations, D={(di,yi)}i=1n0
1: nn0=r×c
2: 1STOPFALSE
3: Obtain p(fD,d~) and p(doptD) using fitted GP             ⊳ Obtain posteriors
4: d*argmind~μ(d~)+σ(d~)                       ⊳ Obtain effective best point
5: f*μ(d*)                             ⊳ Obtain best value of f
6: Calculate αAEI(d~D) for d~D                    ⊳ Compute AEI
7: while  n<N and 1STOP=FALSE  do
8:  d(c+1)argmaxd~DαAEI(d~D)                   ⊳ Obtain next dose
9:  for  i=1,,r  do
10:    Evaluate yi at d(c+1)                      ⊳ Observe outcomes
11:  end for
12:  nn+r; cc+1                       ⊳ Update n,c
13:  DD{(di(c+1),yi)}i=1r                      ⊳ Update data
14:  Obtain p(fD,d~) and p(doptD) using fitted GP           ⊳ Obtain posteriors
15:  d*argmind~μ(d~)+σ(d~)                     ⊳ Obtain effective best point
16:  f*μ(d*)                           ⊳ Obtain best value of f
17:  Calculate αAEI(d~D) for d~D                   ⊳ Compute AEI
18:  Update 1STOP using (4)                      ⊳ Update stopping rule
19: end while
Algorithm 1

Standard Optimization Algorithm

Require:  r patient responses at each of c initial dose combinations, D={(di,yi)}i=1n0
1: nn0=r×c
2: 1STOPFALSE
3: Obtain p(fD,d~) and p(doptD) using fitted GP             ⊳ Obtain posteriors
4: d*argmind~μ(d~)+σ(d~)                       ⊳ Obtain effective best point
5: f*μ(d*)                             ⊳ Obtain best value of f
6: Calculate αAEI(d~D) for d~D                    ⊳ Compute AEI
7: while  n<N and 1STOP=FALSE  do
8:  d(c+1)argmaxd~DαAEI(d~D)                   ⊳ Obtain next dose
9:  for  i=1,,r  do
10:    Evaluate yi at d(c+1)                      ⊳ Observe outcomes
11:  end for
12:  nn+r; cc+1                       ⊳ Update n,c
13:  DD{(di(c+1),yi)}i=1r                      ⊳ Update data
14:  Obtain p(fD,d~) and p(doptD) using fitted GP           ⊳ Obtain posteriors
15:  d*argmind~μ(d~)+σ(d~)                     ⊳ Obtain effective best point
16:  f*μ(d*)                           ⊳ Obtain best value of f
17:  Calculate αAEI(d~D) for d~D                   ⊳ Compute AEI
18:  Update 1STOP using (4)                      ⊳ Update stopping rule
19: end while
Require:  r patient responses at each of c initial dose combinations, D={(di,yi)}i=1n0
1: nn0=r×c
2: 1STOPFALSE
3: Obtain p(fD,d~) and p(doptD) using fitted GP             ⊳ Obtain posteriors
4: d*argmind~μ(d~)+σ(d~)                       ⊳ Obtain effective best point
5: f*μ(d*)                             ⊳ Obtain best value of f
6: Calculate αAEI(d~D) for d~D                    ⊳ Compute AEI
7: while  n<N and 1STOP=FALSE  do
8:  d(c+1)argmaxd~DαAEI(d~D)                   ⊳ Obtain next dose
9:  for  i=1,,r  do
10:    Evaluate yi at d(c+1)                      ⊳ Observe outcomes
11:  end for
12:  nn+r; cc+1                       ⊳ Update n,c
13:  DD{(di(c+1),yi)}i=1r                      ⊳ Update data
14:  Obtain p(fD,d~) and p(doptD) using fitted GP           ⊳ Obtain posteriors
15:  d*argmind~μ(d~)+σ(d~)                     ⊳ Obtain effective best point
16:  f*μ(d*)                           ⊳ Obtain best value of f
17:  Calculate αAEI(d~D) for d~D                   ⊳ Compute AEI
18:  Update 1STOP using (4)                      ⊳ Update stopping rule
19: end while
Close
This Feature Is Available To Subscribers Only

Sign In or Create an Account

Close

This PDF is available to Subscribers Only

View Article Abstract & Purchase Options

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

Close