Biomedical Engineering Reference
In-Depth Information
Route 2
M
2
Route 3
Dt
1
M
1
Route 1
Route 4
M
3
M
4
Dl
2
Route 5
Figure 3.27
Routing plan for Subproblem 3.
manipulation method is first applied without the scheduling algorithm for
aligning droplets. The snapshots are generated by assuming that from the
starting point, each droplet moves one electrode along its pathway in every
snapshot. Note that fluidic constraints might be violated if two droplets
simultaneously move to the next cells. In this case, we can force one of them
to stay in the current location at that snapshot, thereby overriding the con-
straint violation. We calculate the number of manipulation cycles needed
using the grouping-based method. Next, we apply the manipulation method
combined with the scheduling algorithm. Again, we record the number of
droplet-manipulation cycles required. The computation time for the routing
scheduling and the manipulation method for the entire assay is 173 s, on a
Intel Core Duo 2 GHz PC with 2 GB of RAM.
Here, we use Subproblem 3 for illustration. As shown in Figure 3.27, the
routing plan for Subproblem 3 contains five routing pathways (routes). If no
grouping method is used, droplet movements are carried out once per cycle.
The total number of cycles needed equals the number of electrodes on all
the pathways, that is, 60. However, using the grouping-based manipulation
method, the number of cycles is reduced to 53 cycles. Note that no schedul-
ing of droplet routes is used at this time to align droplets. The details of the
manipulations in each cycle are listed in Table 3.4, where the arrows indi-
cate that the target droplet is moved one electrode toward the corresponding
direction. An entry “x” indicates that the droplet stays in the current location
in that snapshot.
In Table 3.4, two droplets are moved simultaneously in several manipula-
tion cycles. This increases the currency of droplet movements. However, due
to the severe misalignment of the snapshots, the number of such concurrent
manipulation cycles is rather limited. Therefore, the reduction of routing
time is quite modest, less than 12%.
