Representatively for one of the most challenging tasks in the industry*

 

TSP = Travelling-Salesman-Problem

The task: find the shortest route for visiting n different cities

Example: visit all 48 state capital cities of the US main-land

 

TSP with 48 cities to visit means

1 * 2 * 3 * 4 * ... * 48 = 48! possibilities to do that

48! = 1.24 * 1061 (that is a 1 with 61 zeros!)

This is more than the world has atoms

And this means a computer would calculate longer than the universe will ever be existing

These problems are therefore called NP-hard

TSPTW = Travelling Salesman Problem with Time Windows

Now we have to visit the different c ities within different time windows

E.g. we have do deliver goods at a specific delivery date

Many problems in the industry can be derived from TSPTW

E.g. order sequencing for make-to-order manufacturer

TSPTW combines an optimization and a scheduling problem

 TSPTW with SuperIntelligence™

So, there is no algorithm hint, which can solve the TSPTW problem analytically and to 100% exactness

The record for solving a TSP problem exactly is by 89.500 cities

The calculation of this 89.500 cities-route took more then 2.000 h

In industry application we can’t wait 2.000 h for a result

SuperIntelligence™ now combines some of the 11 technologies in order to find a pretty good solution within a reasonable timeframe

This combination of technologies run on a massive parallelized hardware in order to accelerate computing time up to real-time

So first the combination of different high-performance algorithms and secondly the parallel execution platform makes to difference to success!

 

Example Combination

1. Determine the cost-function

2. Classify the different productions

3. Implement any restrictions, like delivery dates

4. Minimize the cost-function

5. Simulate in order to visualize all-over costs

 

The SuperIntelligence™ investment in “order sequencing optimization” has redeemed itself after 8.5 month

 (s. ? Benefit)

 

BUSINESS INTELLIGENCE

EMBEDDED SUPERCOMPUTING

HOW IT WORKS

BENEFIT

 

Notes:
* E.g. sequence of production orders can be derived from TSPTW The length of the route is equal to the setup costs between the different production orders = cost-function
The Time Window is the delivery date Additionally the runtime of each order delays the delivery date