Optimisation

Situation

Organisations are striving to make optimal decisions that provide greatest benefit immediately and into the future. Whilst many have a view of their historic performance, few are able to make good predictions about the impact of future decisions on performance and fewer still are able to optimise those future decisions to maximise benefits.

Much optimisation is performed mentally by analysts and managers based on historical performance analysis with often sub-optimal results. Some optimisation is performed through scenario analysis with a necessarily limited number of scenarios. For many business problems proper optimisation brings significantly better results, but it often requires the inclusion of large quantities of input data and a significant number of complex business rules. Organisations often believe building optimisations to support the required analysis of this data and these business rules is too complex to be cost-effective or provide reliable answers.

Solutions

Optadian has extensive experience with delivering practical optimisation solutions for clients that support key business decision making. We have worked with clients to

• analyse key business decisions
• identify and analyse the data behind the decisions
• determine the business rules around the decision
• identify the optimisation that needs to be perforated
• develop the logic surrounding the optimisation
• develop the IT infrastructure to support the optimisation
• work with the client to integrate the optimisation into the business process
• train key staff to be able to use, manage and enhance the optimisation

We have developed solutions for clients using toolsets including Frontline Systems' Analytic Solver Platform (www.solver.com) and IBM's ILOG CPLEX (www.ibm.com) as well as using the optimisation capabilities within simulation tools such as Powersim Studio from Powersim Software (www.powersim.com).

Optadian has worked with IBM to provide input over the future development of their ILOG CPLEX product and has worked with them to improve efficiency of some areas of their product.

Example Credentials

Optadian has implemented optimisations for a range of clients which have enabled them to develop optimal plans taking into account their understanding of the market and business rules.

European National Transmission and Distribution Gas Utility

Optadian architected, designed and implemented an optimisation solution to predict the capacity purchases across the entire national market for the transmission network given changing regulatory and market scenarios. The optimisation included inter-company trading. The optimisation was used to simulate a near perfect market where overall market costs are minimised, enabling the client to understand the likely capacities required at different points in the future and the revenues associated with these.

Optadian initially assessed a prior optimisation implementation which had limited functionality and poor performance, performing extensive stress and performance tests on it. After determining that the existing solution architecture and software toolset were incapable of supporting the business requirements, Optadian rearchitected and fully redesigned the solution, incorporating all the new business functionality such as inter-company trading and changing the toolset to the higher performance IBM ILOG CPLEX.

The performance of the solution was improved by more than an order of magnitude despite having significantly more functionality and input data included. Additionally Optadian rearchitected the output data handling to incorporate a multi user Microsoft SQL Server database so that a wide range of users (with suitable permissions) could access the output results and view them.

Using the solution the client is able to make accurate predictions for the purchases and trades in the capacity market, determine the revenues that will occur and use this to justify different courses of action with the regulator.

Global High Value Commodity Mining, Marketing and Sales Monopoly

Optadian staff architected, designed and implemented a multi-user sales allocation optimisation solution for the sales staff of the monopoly company to ensure that their sales were directed towards clients in a way that provided maximum strategic return on sales whilst remaining compliant with European Union competition legislation.

The solution was developed as a custom Java development that incorporated IBM's ILOG CPLEX optimisation toolset.

Using the solution the client was able to ensure that their $5bn+ per year of sales was correctly targeted to maximise strategic return and that they were compliant with European regulations.

In addition to the main optimisation solution, an additional proof of concept solution was developed for price point constraint management using a digraph (directed graph), a technique later used by companies such as Google and Twitter to visualise large quantities of data. This had the potential to transform this part of the business, where previously the recalculation and re-optimisation of the 80,000 price points and the removal of inconsistencies between them would take days of painstaking work, to a sub-second graphically driven process. Whilst the proof of concept was too advanced for the organisation at the time, it was picked up again several years later by them and developed into a working solution.