Many Planning Options in Parallel

Dynamic programming is a standard tool for capacity expansion planning.
Expansion options are gradually pruned until an optimal solution is achieved.
Pomegranate parallelization results from sophisticated
relaxation of dependencies. The number of states relevant to the
optimal solution is reduced by progressing in parallel along time
as seen in the following illustration.

Many alternatives in PARALLEL, rapidly reduce an initially large number of expansion options.

Significant and Consistent Speedups

Pomegranate, after extensive tests, has been in production run
for years in IEC - Israel Electric Corporation. It has shown consistent
and significant speedups for a variety of input data, in diverse platforms.
Measured results - for 2 datasets: GREEN and YELLOW - in the figure,
show their quasi-linear characteristic, compared with the linear values in RED .
Results are obtained in short times, instead of long hours.

SPEEDUP

Number of processors

Measured values

Pomegranate - a Computing Space compatible product

Uni-processor and multi-processor runs use the same input files
in a standard format and produce the same output files.

Pomegranate runs upon the platforms in the market, possibly in an heterogeneous environment.
Platform independence is guaranteed since Pomegranate was developed using PCSpace.
Therefore it is Computing Space compatible.