5.5 Developments and Change within Management Information Systems

5.5 Developments and Change within Management Information Systems

Introduction
Once a management information system is installed it enters the maintenance phase of the system lifecycle and will change with time.
In addition to perfective and corrective maintenance, a lot of time may be spend in adaptive maintenance to cope with changing requirements.
Businesses rarely remain static because they are part of their evolving environment and must change with it. These changes may be as simple as developing new products or changing the company’s entire focus.
Such changes are often influenced by the MIS and guided by decision support systems and expert systems.
A new MIS is not guaranteed to succeed. It can fail for a variety of reasons, such as weaknesses in the system itself or because of the humans involved in it.
By the end of the section you should be able to describe:
- Why an information system might fail;
- The changes to a company when a new information system is introduced;
- The main features, uses and benefits of decision support systems;
- The main features, uses and benefits of expert systems.

5.5.2 Introducing a new Information System
A new system often does not run smoothly due to either technical or personnel glitches.

When introducing a new information system the following points should be addressed in order to ease the experience for employees. Their precise order depends on the method of implementation adopted and other circumstances unique to the company. But how they are handled will directly affect the attitude employees have to the new system.

Data Conversion: If data is not properly imported, or if users have to go through many ‘hoops’ to access data from the old system because of inefficient design, they may become frustrated.
New Work Practices: The way that people do their jobs will change in some way. Some people may be glad to be rid of mundane tasks while others may want to hold onto them because they are uncomfortable with them. Employees may view a new system as either the solution to their problems, or as a needless replacement for system that worked.
Allowing the users to continue to access old software: For example, a company moving from an office package to another may decide to give users six months to migrate at their own pace from one to the other while training is organised. However, some users may decide they prefer the old package and may stubbornly hold onto it. Other companies may do the opposite and force users to change systems overnight. If the new system is not satisfactory some staff may stop working and complain that they cannot use the system because it does not work.
Training: Busy employees may resent training, seeing it as a waste of time that disrupts thr work they have to do. There are those who would rather use a system inefficiently than attend training because of the feeling of ‘doing something’, as opposed to the feeling of wasting time. This negative attitude exists because some training seminars are actually of questionable value because they are poorly presented or take longer than necessary. Training must be presented to employees as something that will benefit both them and the company.

It thus can be seen that the attitude of employees contributes to a system’s failure or success. However, if employees are treated in a sensible and respectable manner, many potential problems can be avoided.
A system’s success is also determined by the external and internal factors considered in the previous section. At all times a system must remain relevant to the company’s mission, environment and people . If it fails to adapt, a system that is acceptable may soon be rendered useless. Systems may fail even before they are implemented. Poor analysis can lead to many errors, an unworkable design, obsolete features, or even the project’s cancellation. A system that does not do what it is expected or required to do may have lasting effects on a business.
Delivering a successful system is therefore much more than good design and programming.


5.5.3 Decision Support Systems (DSS)
A decision support system is a set of integrated tools designed to help in problem solving, such as scheduling work activities, allocating resources and forecasting future trends. A DSS can operate as a stand-alone tool, though it is more often integrated with existing transaction processing systems and/or MIS. DSS are basically problem solving tools that analyse data gathered by other systems and combine this data with decision making models to produce information to help the user solve problems.

A DSS has a number of distinct features that make it much more than a ‘powerful MIS’,
- It brings together data and mathematical models to support human judgement.
- It supports several interdependent decisions, by modelling the impact that differing problems have on each other.
- It supports a wide variety of decision making processes and styles.
- It assists decision making within dynamic business conditions.
- It supports AD HOC queries.

A DSS can assist with several different types of problem including:
- Independent Problems: Problems that are completely separate from each other. In this case, the goal is to find the best solution to a single problem.
- Interrelated Problems: Problems that affect each other. The goal is to find the best overall set of solutions, not just the solution to independent problems.
- Organisational Problems: These are problems that span a number of departments within an organisation and may affect the organisation as a whole.

A DSS has three main components:
- The DBMS stores internal and external data that are analysed by the DSS. The MIS accesses the same DBMS.
- Model Management System – takes input data, perform some sort of computation upon it and deliver output that is often in the form of a forecast. There are many different kinds of model. Statistical models are used to analyse statistics, such as production rates or sales figures. Financial and accounting models assess the financial implications of different courses of action including ‘optimistic’, ‘pessimistic’ and ‘realistic’ scenarios of what may happen.For example, in a optimistic scenario, sales may be 30% above the expected rate and in a pessimistic scenario, sales may be 60% less than expected. Production models perform functions such as calculating the number and type of machines needed, amount of raw materials required and their rates of consumption. Marketing models are used to aid decisions such as locating new stores, pricing products and forecasting sales. Finally, human resource models help managers to make decisions that involve personnel, including planning numbers of workers needed, assessing training needs, maintaining a skills inventory and assessing implementation of government rules and regulations.
- Support Tools allows a user to interact with the system. They include pull down menus, online help, user interfaces and tools for graphical analysis.

Although we have seen how a DSS is composed and the types of problem it can deal with, the underlying function of a DSS is fivefold.
- Model building, i.e. identifying appropriate models for solving a given problem. This involves analysing input variables, relationships between the variables, assumptions made about the problem and constraints on the problem.
- ‘What if’ Analysis is used to assess what a changing variable will have on operations, for example ‘what if interest rates rise?’, ‘what effect will rising oil prices have on manufacturing costs’, or ‘how much will demand increase if we reduce prices’.
- Goal seeking works in the opposite way. A desired goal is entered and analysed, allowing decision makers to determine the input values needed to achieve this. For example, a company may want to increase profits by 10% but seek to do this without forced redundancies or hefty price rises.
- Risk analysis assesses the uncertainties of different courses of action. Probability statistics are used to evaluate these risks.
- Graphical Analysis allows data and information to be viewed as graphs and charts.

5.5.4 Expert Systems
An expert system is an application that performs some task that a human expert would otherwise perform and does so at, or near, the skill level of the human expert. Such as system is given the knowledge a human ‘expert’ would have in a specialist field and, based on that knowledge, it makes recommendations. While a DSS makes recommendations that humans are expected to discuss, evaluate, and query further, an expert system is expected to give the correct answer without the need for discussion. This means that it is suitable only for certain applications where rules can be clearly defined.
Some human experts may use an expert system to give them a ‘second opinion’. For example, a doctor may use such a system to analyse x-ray images. Alternatively, a human who is not an expert may use such a system to help them make decisions. For example, an expert system could carry out fault diagnosis in a machine before repair details are passed to the human expert. This approach is efficient because the human expert does not have to waste time making a diagnosis when the computer can perform this task. Expert systems have been developed in many fields, including medicine and law.
Many expert systems are really a branch of artificial intelligence (AI) and like many AI systems seek to model some aspect of human reasoning. An expert system must be capable of taking the same inputs as a human expert and being correct in its outputs at least as often as the human. And, like any human expert, it needs the ability to learn from experience.
Alan Turning, considered by many to be the father of artificial intelligence, proposed what is now called the turing test. This states that a computer can be considered intelligent if, following a conversation with a human via a remote terminal, the human cannot tell if they were talking with a computer or a human.
Although no system has comprehensively passed the turing test, some are considered to be as effective as humans in limited areas, such as specialist legal or game playing systems, because they consistently give the same output as a human would.

An expert system has the following features:
- It is limited to a certain area of knowledge (domain)
- It is based around rules, facts and principles.
- It can deal with ‘fuzzy logic’ (instead of plain yes/no, it can process a third value – ‘don’t know’ or ‘maybe’).
- The system’s reasoning can be explained to the user.
- It is capable of learning from experience.

An expert system has three main components.
- A knowledge base which is a store of facts, rules and principles from a given field. Knowledge representation is the process of translating knowledge into a form that can be programmed into the expert system.
- The interface engine solves a problem by applying the rules and knowledge already in the system to the facts that are entered concerning the problem.
- The user interface which includes menus, graphics and facilities for explaining the system’s reasoning.