Frequently Asked Questions

Please submit your questions about any of our product, services, and training programs to: 


Q. What is an expert system?

An expert system is a computer program that is designed to emulate the decision making of a human expert. While an expert system is not meant to replace the scientist, it does provide expert knowledge in an organized system capable of continued growth.

Q. How can a pharmaceutical scientists utilize an expert system?

Within the pharmaceutical industry, there are many uses for expert systems. For example, an expert system can provide formulation and process justification, formulation and process optimization, and training. Each of these applications is a tool that will make the job of a human expert easier and more efficient.

Q. What is the development plan of PTI for customized systems?

  • Establishment of a project team: A project team (task force) is formed in order to develop this project. PTI's representative(s) meets several times to develop a strategy for the implementation of the proposed project. Overall responsibility of the client is to provide the "domain" knowledge in these meetings. PTI then determines which software to use in order to transform that knowledge into a coded program for the development of the expert system

  • Feasibility study: The project team determines the motivation of the company to develop an expert system (See Appendix 1-a given at the bottom of this page), the features of each modules of the system described in specific projects (See products such as SPRAYex, FILMex etc). The project team also uses the general feasibility guidelines whenever it is applicable (See Appendix 1-b,c,d).

  • Development of the flow charts: The project team develops flow charts for each specific problem.

  • Development of the database (tables): All of the appropriate data are converted to (Access) databases.

  • Acquisition of the knowledge: The project team meets several times to develop rules for each specific problem or critical steps involved in the development of the particular formulation and process.

  • Development of the modules: PTI applies appropriate programming techniques and develops each module according to the specifications determined by the project team.

  • Testing the modules and development of the prototype: PTI uses case studies with known results to test the ability of the rules, databases, and programming to perform properly. The client provides the relevant case studies.

  • Implementation, testing and trouble shooting of the final program: PTI uses known case studies as well as untested materials and parameters to verify the proper operation of the program and to troubleshoot any additional problems identified.

  • Training of users: A user acceptance questionnaire is be utilized during the implementation of the program.

  • Maintenance of the program: PTI provides free service for three months after the implementation of the program to trouble shoot the problems encountered during the initial utilization of the program. During this maintenance period, PTI also trains the users of the program. If the client purchases an additional yearly renewable maintenance agreement, PTI continues to provide service to trouble shoot the problems encountered after the three month period. The yearly maintenance program also covers annual training for new and advanced users of the expert systems.

  • Updates of the program: Subject to separate negotiations.

Q. Which program techniques does PTI use in the expert system development?

PTI integrates a number of knowledge representation techniques in order to maximize the performance of the expert system. Some of these techniques are briefly described below:

  • Induction Systems: One simple way of constructing a small expert system is to represent the experts' knowledge in a table of attributes and use of an induction algorithm to convert the experts' knowledge into a decision tree.

  • Rule Based Systems: The most common way of representing knowledge and handling inference and control is found in rule-based systems. This approach represents knowledge as statements and rules and uses forward and backward chaining to handle the inference and control. Two types of rule based systems are available. The simpler rule-based systems group all the rules together in one set and then examine them all at once. The more complex or structured rule-based systems divide the rules into subsets, arrange the subsets into a tree, and then examine them according to some search strategy. A rule is an IF-THEN-ELSE-BECAUSE construct that links statements together in order to facilitate inference.

    • Control by Backward Chaining: Backward chaining is by far the most common strategy used in the simple rule systems. Backward chaining is a term used to describe running the rules in a "goal driven" way. A goal is an attribute for which the system is trying to establish a value. In backward chaining, if a piece of information is needed, the program will automatically check all the rules to see if there is a rule that could provide the needed information. The program (inference engine) will then "chain" to this new rule before completing the first rule. This new rule may require information that can be found in yet another rule. The program will then again automatically test this new rule. The logic of why the information is needed goes backward through the chain of rules.

    • Control by Forward Chaining: Forward chaining is a "data driven" way to run the rules. In backward chaining, there is always a goal to be satisfied and a specific reason why rules are tested. In pure forward chaining, rules are simply tested in the order they occur based on available data. If information is needed, other rules are NOT invoked - instead, the user is asked for information. Consequently, forward chaining systems are dependent on rule order. However, since time is not spent determining whether the information can be derived, forward chaining is much faster.

    • Control by a Hybrid Between Backward and Forward Chaining: In this case, the basic approach is data driven, but information needed by rules is derived through backward chaining.

    • Confidence Factors: Confidence factors, or certainty factors, refer to a numerical weight given to a fact or a relationship to indicate the confidence one has in that fact or relationship. In a typical expert system programming language, there are several ways of handling uncertain data.

  • Fuzzy Logic: Fuzzy logic is a powerful technique that enables the developer to built expert systems that more closely reflect the real world.

  • Case-based Reasoning: To solve a problem it searches and finds a similar problem that was solved in the past and adapts the old solution to solve the new problem.

  • Neural Networks: Neural Networks, also known as artificial neural networks, are a software approach to emulate the learning behavior of living nerve cells in animal physiology. The basic processing unit is the neuron, which takes one or more inputs and produces an output.

  • Front-end Programming: PTI uses a front-end programming technology and integrates as many artificial intelligence techniques as possible in order to enhance the power of the ultimate software product. In fact, this is absolutely necessary if a successful system in the area of pharmaceutical technology is desired to be developed. For example, a neural network can recognize patterns in historical data and can also modify itself through time due to changes in conditions external to the expert system, allowing the system to be a dynamic model and self-modifying. However, neural network software has several drawbacks including the lack of a runtime interface, access to external data sources (i.e. databases and spreadsheets) and explicit control; all of which can be provided in a rule-based expert system inference. There are tremendous benefits to combining neural networks with a rule-based system in application development. Professional rule-based systems alone are not best at automated learning or recognizing patterns in large amounts of data. This gap in expert systems is filled by neural networks

Q.  What are the hardware, software and operating system requirements?

  • The PTI expert systems runs on the IBM compatible PCs with Windows 2000/XP operating systems.

  • A hard-lock is provided free of charge by PTI for the authorized use of the software. 

  • Network applications can also be negotiated with PTI. 


Improved Productivity:

  • Better Decisions: The system is expected to be capable of improving the quality of decisions.

  • Disseminates Expertise: The system is expected to provide expertise to locations within the organization where this capability is lacking.

  • Faster Decisions: The system is expected to reduce the time to reach a decision.

  • Other: The system is expected to ______________________.

Lower Costs:

  • Reduces Labor Costs: The system is expected to reduce labor costs by allowing a time-consuming task to be completed quickly or acts in place of a highly paid expert.

  • Improves Material Use: The system is expected to improve the use of materials during manufacturing.

  • Other: The system is expected to ______________________

Improved Quality:

  • Superior Product: The system is expected to improve the quality of the final product.

  • Superior Service: The system is expected to improve the quality of services supplied by the organization.

  • Provides Training: The system is expected to provide training to personnel that improves their work activities,

  • Other: The system is expected to ______________________.

Improved Image:

  • Innovator: The system is expected to improve the organizations image as a leader and innovator.

  • Other: The system is expected to _____________________.







Expert knowledge needed




Problem-solving steps are definable




Symbolic knowledge used




Heuristics used




Problem is solvable




Successful systems exists




Problem is well focused




Problem is reasonable complex




Problem is stable




Uncertain or incomplete knowledge








Solution more of a recommendation









Problem Solving Feasibility = Total Score / Total Weight 










Can communicate knowledge.




Can devote time.












Good communication skills




Can match problem to software




Has ES programming skills




Can devote time







Can devote time




Receptive to change













Support project




Receptive to change




Not skeptic




Reasonable expectations




Understands objectives









People Feasibility = Total Score / Total Weight 







System can be introduced easily




System can be maintained




System not on a critical path




System can be integrated




Training is available









Deployment Feasibility = Total Score / Total Weight






Ease of use



Starting the system



Obtaining explanation



Help facilities



Interface techniques



Exiting the system



Nature of questions



Clarity of terms



Answers complete



Clarity of questions



Nature of explanations



WHY explanations



HOW explanations



Presentation of results



Easy to follow






System utilities



Easy to access






General considerations



Speed of system



System is useful



Provide any general comments about the system:



  • Expert System: A computer program designed to model the problem-solving behavior of a human expert.

  • Heuristic: Knowledge, often expressed as a rule of thumb, that guides the search process.

  • Domain Expert: A person who possesses the skill and knowledge to solve a specific problem in a manner superior to others.

  • Problem Solving: The process of seeking a solution to a given problem.

  • Symbol: An alphanumeric pattern that represents some object characteristics or event of a problem.

  • Symbolic programming: Manipulating symbols that represents objects and their relationships.

Please submit your questions to: