1916 - 2001 Extensions 1. Intelligence • Is there a Problem ?? • Should I get dressed today ?? Simon’s Model of Decision Making: 2. Design • What Alternatives Are there ?? • Should I wear Black Socks? Blue Socks? Red Socks? No Socks?? 3. Choice • OK – The Red Socks! 4. Implementation • Put on the socks already! 5. Review • How do they Look? Was it a good choice?
But what steps do actually we take when making a decision? Simon’s Model corresponds to the Scientific Approach, which indicates the specific tasks needed AND also corresponds to the Systems Development Life Cyle (SDLC): Simon System’s Approach SDLC System Investigation Intelligence Define Problem Design Develop Alternatives System Analysis System Design Choice Select Solution Design Solution Implement Solution Implementation Imple-mentation Maintenance Review
Consider The Following Situation Grandpa John passed away at the ripe old age of 102. He left a pleasant surprise to his 5 great-grandchildren. He had dabbled in the stock market and accumulated enough to leave $10,000,000 after taxes to his great-grandchildren. The will contained one stipulation – The money could not be divided for 20 years. It was the responsibility of the great-grandchildren to manage the investment collectively.
From Simon’s Perspective: Is there a problem?? • Only if the Great-Grandchildren want to maximize their Inheritance From The System’s Approach: Define the Problem • We need to find an optimal portfolio that would maximize return on investment over a 20 year period What do we do ??? Ennumerate/Develop Alternatives • Nothing • Go to Las Vegas • Put the money in savings • Buy Lottery Tickets • Put the money in Bonds • Give the money to Charity • Put the money in Stocks • Manage the money collectively • Invest in Property • Cede control to one member • Buy a Racehorse • Hire a Financial manager • Consult a Fortune Teller • Develop a Portfolio of Investments
This first step is called Brainstorming: • focus on a specific problem • deliberately come up with as many unusual solutions as possible and push the ideas as far as possible. • Brainstormers should not only come up with new ideas in a brainstorming session, but should also 'spark off' from associations with other people's ideas and develop other peoples ideas. • Ideas MUSTNOT be criticized or evaluated during the brainstorming session. • Criticism introduces an element of risk for a group member in putting forward an idea. • Criticism stifles creativity and cripples the free running nature of a good brainstorming session
Brainstorming may be individual or done in groups • Individual brainstorming tends to produce a wider range of ideas, but tends not to develop the ideas as effectively, as individuals run up against problems they can’t solve. • Group brainstorming develops ideas more deeply and effectively since another person's creativity and experience can be used to break ‘creative blocks’. • Group brainstorming tends to produce fewer ideas (as time is spent developing ideas in depth) and can lead to the suppression of creative but quiet people by loud and uncreative ones. • Individual and group brainstorming can be mixed, perhaps by defining a problem, and then letting team members initially come up with a wide range of possibly shallow solutions which can then be enhanced and developed by group brainstorming.
So now that we have decided on all options, we can choose one, right ??? It wouldn’t make much sense to choose one without analyzing each of them. ??? HOW ??? That is NOT so easy. There are four basic models of Decision making: • The Rational Model • The Political Model • The Process Model • The Garbage Can Model Which one to use depends on what you are trying to achieve
The Rational Model: • based upon an economic view of decision making. • It is grounded on goals/objectives, alternatives, consequences and optimality. • assumes that complete information regarding the decision to be made is available and one correct conception of a problem, or decision to be made can be determined. • assumes that the decision-makers consistently assess the advantages and disadvantages of any alternatives with goals and objectives in mind. • assumes that the decision-makers then evaluate the consequences of selecting or not selecting each alternative. • assumes that the alternative that provides the maximum utility (i.e., the optimal choice) will be selected. ??? Isn’t this how all decisions are made ??? Not Really
The Political Model: • considers the preconceived notions that decision-makers bring to the table in the decision process. • the individuals involved do not accomplish the decision task through rational choice in regard to objectives. • The decision makers are motivated by and act on their own needs and perceptions. • involves a cycle of bargaining among the decision makers in order for each one to try to get his or her perspective to be the one of choice. • More specifically, this process involves each decision-maker trying to sway powerful people within the situation to adopt his or her viewpoint and influence the remaining decision-makers ??? Who would ever use that ??? Ever heard of the US Congress?
The Process Model: If A then B • decisions are made based upon standard operating procedures, or pre-established guidelines within the organization. • the organization of past, present, and future events, as well as conformity, are integral parts to this model because they can be used as a consistent foundation for decision making. • Conformity is an integral part of the process model since it is the means by which doubt, or incertitude is dealt with during the decision task. • If decision makers are uncertain as to the potential effectiveness or the results of a decision, they conform to the pre-established standard. ??? Where would this be used ??? Ever heard of unions?? Civil Service ?? Government Agencies ??
The Garbage Can Model: • Used for judgment tasks: participant involvement participants fluctuates in the amount of time and effort given, & choices are inconsistent and not well defined • an opportunity to make a decision is described as a garbage can into which many types of problems and solutions are dropped independently of each other as these problems and solutions are generated. • The problems, solutions and decision makers are not necessarily related to each other. • The components are the combination of options available at a given time, the combination of problems, the combination of solutions needing problems, and the external demands on the decision makers ??? How would this work ??? Ever see what goes on here in the college at the end of the year if departments have excess funds to spend (or lose them if they don’t)?
A DSS is intended to: • Provide Support for individual (or Group), Ad hoc (impromptu), decision making ?? How do Decision Support Systems (DSS) fit in?? • Provide Support primarily for semi-structured or unstructured decision making (unlike standard Management Information Systems) • Combine data, analytical models and tools, and software into a single system • Note that a DSS is NOT intended to: • Provide the solution (it is only tool) • Be used over and over again (It was designed for unique decision making) • Always used the same analytical models and tools (The decision maker chooses the models and tools based on the problem at hand)
Is there a theory to support the use of Decision Support Systems? Yes, And No -- NO - There is not really a theory of Decision Support Systems YES - There are a series of propositions (by our old friend Simon), that lay the foundations for DSS Usage
(First from his 1945 Text on Administrative Behavior) Proposition 1: “Information stored in computers can increase human rationality if it accessible when needed for making decisions” Simon’s Propositions Proposition 2: “Where a particular item of knowledge is needed repeatedly in decision, the organization can anticipate this need and, by providing the individual with this knowledge prior to decision, can extend his area of rationality. This is particularly important when there are time limits on decisions.” Proposition 3: “Specialization of decision-making functions is largely dependent upon the possibility of developing adequate channels of communication to and from decision centers”
(Next from his 1973 Text on Applying Information Technology) Proposition 4: In a post-industrial society, the central problem is not how to organize to produce efficiently but how to organize to make decisions -- that is, to process information. Simon’s Propositions Proposition 5: For information processing, division of labor means breaking-up the decisions to be made into relatively independent subsystems, each one designed with only minimal concern for its interactions with the others. Proposition 6: The key to the successful information systems design lies in matching the technology to the limits of the human capability. In general, an additional component (man or machine) of an information-processing system will improve the system's performance only if: 1. It conserves attention instead of making additional demands on attention 2. It incorporates processes that automatically select and filter information for subsequent transmission; 3. It incorporates analytical models that not only store and retrieve information, but also assist in solving problems, evaluating solutions, and making decisions.
What does all that mean? • It outlines the basic components needed for a DSS AND how they should be constructed. SO?? • Let’s Consider each of the propositions separately:
TPS MIS Corporate Database Proposition 1: “Information stored in computers can increase human rationality if it accessible when needed for making decisions” The implication here is that there MUST be a database of facts available prior to making a decision
External Databases Policies TPS MIS Corporate Database Business Unit Databases Proposition 2: “Where a particular item of knowledge is needed repeatedly in decision, the organization can anticipate this need and, by providing the individual with this knowledge prior to decision, can extend his area of rationality”. The implication here is that unanticipated data and facts must be made available as the need arises (It must be flexible)
User Interfaces Corporate Database Proposition 3: “Specialization of decision-making functions is largely dependent upon the possibility of developing adequate channels of communication to and from decision centers” The implication here is that information must be shared between the decision makers: GDSS
User Interfaces Models & Methods Tools & Techniques Corporate Database Proposition 4: In a post-industrial society, the central problem is not how to organize to produce efficiently but how to organize to make decisions -- that is, to process information. The implication here is that there must be DSS software that will consolidate models and analytical tools: • Sensitivity Analysis • Statistical Models • Spreadsheets • Regression Analysis • Forecasting • Data Mining • Time Series • Pattern Recognition • Optimization • Fuzzy Logic Models • LP, IP, Optimal Flow • AI Shells
Manufacturing Product Design User Interfaces Models & Methods Tools & Techniques Corporate Database • Regression Analysis • SensitivityAnalysis • Statistical Models • Spreadsheets • Forecasting • Data Mining • Time Series • Pattern Recognition New Product • Optimization • Fuzzy Logic Models • LP, IP, Optimal Flow • AI Shells Marketing Finance Proposition 5: For information processing, division of labor means breaking-up the decisions to be made into relatively independent subsystems, each one designed with only minimal concern for its interactions with the others. The implication here is that the DSS will flexible enough to deal with various problems
User Interfaces Models & Methods Tools & Techniques Corporate Database • Regression Analysis • SensitivityAnalysis • Statistical Models • Spreadsheets • Forecasting • Data Mining • Time Series • Pattern Recognition • Optimization • Fuzzy Logic Models • LP, IP, Optimal Flow • AI Shells Proposition 6: The key to the successful information systems design lies in matching the technology to the limits of the human capability. The implication here is that the DSS will support the individual’s (or group’s) manner of thinking and preferences
GDSS: Group Decision Support Systems We might first use it to help generate alternatives: ?? How could a DSS help in our Inheritance situation?? • same-time-same-place conferences (decision rooms) are oriented toward business meetings and decision making. • All DSS software and brainstorming and groupware (e.g., Lotus-Notes) included • participants simultaneously type in ideas • The GDSS sorts the ideas, and the participants then vote or comment on which ideas they like or they dislike. ??? Advantages ??? • Time Saving: The GDSS Organizes and sorts all inputs • Anonymity: Users identities not given, and hence free to propose anything • No one can dominate the meeting: No fear of repercussions
Group members generally have many biases (REMEMBER: Not everyone is using the Rational Model) The DOD’s Decision Room Of Course: ?? Are there any Disadvantages ?? • Groupthink (buy-in by group members without thinking) can lead to bad decisions. • There may be many alternatives to consider • There are often several conflicting objectives • The result of a business decision usually materializes in the future. No one is a perfect predictor of the future, especially in the long run. • Most (if not all) decisions involve risk. Different people have different attitudes toward risk. • There may not be sufficient information to make an intelligent decision. • There may be TOO MUCH information (information Overload). • Gathering information and analyzing the problem takes time and is expensive. It is difficult to determine when to stop this and make a decision.
Identify the conflicting objectives. • Identify the uncertainties (Risk Rating). • Identify the alternative courses of action (can they be combined?) and the criteria that should be considered in making a decision. ?? What does the group have to decide on ?? • What are the possible results of the decision? Why might the results be difficult to predict? • Can the decision be changed if the economic environment changes? How hard or easy would it be to change the decision once it is made? What could it cost? ?? How does the GDSS Help with that ?? • It can’t help in all situations, but it can help with some: • It can order objectives by ratings • It can Provide summaries of activities • It can Provide Analysis tools for the group/individual members to use
User Interfaces Models & Methods Tools & Techniques Corporate Database • Regression Analysis • SensitivityAnalysis • Statistical Models • Spreadsheets • Forecasting • Data Mining • Time Series • Pattern Recognition • Optimization • Fuzzy Logic Models • LP, IP, Optimal Flow • AI Shells As we noted earlier, that depends, but the more the merrier ?? What Analytical Tools are available ?? ?? But which should we Choose ?? That depends on the decision maker: • A DSS SUPPORTS, NOT decides, the decision to be made • A DSS SUPPORTS the Decision Maker’s style of thinking (it does NOT change it)
There are three ‘Basic’ Classifications of DSSs: • Model-Driven DSS • The DM selects the Tool to be used (e.g., Spreadsheed, Optimization) ?? Are All DSS the Same ?? • The DM selects the Method to be used (e.g., Sensitivity Analysis, Maximize) • The DM Uses various parameters (e.g., Interest Rate Intervals, Constraints) • Data-Driven DSS • Due to Technology Advances, Large amounts of data can be analyzed • The DM can compare various datasets • “Compare Sales of Product A in Regions 1, 2 and 3, relative to operating costs” • Often done using External Database and On-Line Analytical Processing (OLAP) systems • Often used in conjunction with datamining
There are three ‘Basic’ Classifications of DSSs: • Discovery-Driven DSS (Datamining) • The DSS looks for patterns and relationships in large data sets due to: ?? Are All DSS the Same ?? • Associations between occurrences which can be linked to an event (e.g., Overcoat sales) • Sequences in events over time (e.g., increases in new home sales lead to a 30% increase in the number of refrigerators sold) • Classification of patterns (e.g., State employees were less likely to purchase life insurance than non-state employees) • Clustering of groups that have not been defined (e.g., Females, The Elderly, and Europeans might be more likely to purchase a certain item) • Uses statistical analysis tools such as: • Neural Networks • Expert Systems • Fuzzy Logic • GIS • Genetic Algorithms • Other AI Systems
????????????? Any Questions (Please !!!) ?????????????