Polysar (Pricing Advisory System)
Polysar (Pricing Advisory System)
Customer Polysar
(Fribourg, Switzerland)
Objectives
Development Status Developed and delivered a prototype in KEE on a TI-Explorer, during 1988

Highlighted features

The system was developed for Polysar, a Canadian petrochemical manufacturer. Polysar has its European headquarters in Fribourg, Switzerland, and is a UNISYS customer. As a large, international corporation, Polysar faces a constant challenge in strategically pricing its products in the variety of countries sand markets in which is does business. This problem has always been a difficult one, but never more so than in the increasingly competitive global marketplace. In addition to increased competition, the opening p of non-traditional markets (in the developing world, for example) means that the complexity and variety of markets has also increased. The goal of the expert system for Polysar is to aid in determining an effective pricing structure for their products.

Why an expert system for Polysar?

1. There are few experts, but many users.

A human expert has only so much time to advise her colleagues who need that advice. One of the main advantages to expert systems is that the knowledge that experts can only develop with years of experience can be embodied in a computer system that can be accessed by many users. Human experts that are in great demand can now be "consulted", through the expert system, much more widely.

2. There is an evolving knowledge domain.

The global marketplace is constantly in flux, and it is not only the trade laws that change, but the strategies for pricing in the various world markets. A well developed expert system reflects explicitly in its structure the real structure and strategy of the application. Therefore, it can be readily adapted, and can evolve over time as the application evolves.

3. Decision must be made quickly.

This relates to the first reason. As the demand for expert opinions has gone up, the time available to make decisions about pricing has shrunk. This intensifies the desire to have this expert knowledge available on-line.

4. Training.

As new people are brought in who must make pricing decisions, the can learn from the expert system how the expert would evaluate a certain pricing situation. Hypothetical situations can be developed for which these new people must use the expert system to help develop a pricing strategy. This kind of training is flexible, and most importantly, up-to-date.

The modalities which Polysar experts consider in developing pricing strategies are:

  • customer country
  • customer type
  • product application
  • product type
  • market situation
  • competition

This is directly reflected in the structure of the expert system, which has software structures representing customers, countries and markets, and products. The pricing strategies of the Polysar experts are reflected in the rule-base of the system.

We recommended an implementation based on the Explorer Workstation running the KEE artificial intelligence application development system. KEE provides not only sophisticated structuring of data and rule bases, but has an advanced graphical user interface. As Polysar maintains a database of currency rates and customer data on a UNISYS 1100, we also recommended using a PW2 for down-loading data from the 1100 into the Explorer application.


Sample screens

Fig.1- End-user Interface: Interrogation sheet
Fig.1- End-user Interface: Interrogation sheet



Return to the index