Download presentation
Presentation is loading. Please wait.
1
Issues, Trends and Strategies for Computer Systems Management UMUC Graduate School of Management and Technology Chapter 10. Complex Decisions and Artificial Intelligence
2
2 Agenda Specialized Problems Diagnostic Problems Speed Consistency Training DSS and ES Building Expert Systems Knowledge Base Knowledge Engineers Creating an ES Reasoning Limitations of Expert Systems Additional Specialized Problems Pattern Recognition/Neural Nets Voice & Speech Recognition Language Comprehension Massively Parallel Computers Robotics and Motion Statistics, Uncertainty, Fuzzy Logic DSS, ES and AI Intelligence Object Orientation
3
3 Computer analysis of data and model. Decision Operations Tactics Strategy Neural network Company Complex Decisions & Artificial Intelligence
4
4 Specialized Problems Diagnostics Speed Consistency Training Case-based reasoning
5
5 Expert System Knowledge Base Symbolic & Numeric Knowledge If income > 20,000 or expenses < 3000 and good credit history or... Then 10% chance of default Rules Expert decisions made by non-experts Expert
6
6 Expert System Example Bears by ExSys Link: http://www.exsysinfo.com/ Type of Food: Fish Berries...
7
7 DSS and ES
8
8 ES Example: bank loan Welcome to the Loan Evaluation System. What is the purpose of the loan? car How much money will be loaned? 10,000 For how many years? 5 The current interest rate is 10%. The payment will be $212.47 per month. What is the annual income? 24,000 What is the total monthly payments of other loans? Why? Because the payment is more than 10% of the monthly income. What is the total monthly payments of other loans? 50.00 The loan should be approved, there is only a 2% chance of default. Forward Chaining
9
9 Payments < 10% monthly income? Other loans total < 30% monthly income? Credit History Job Stability Approve the loan Deny the loan No Yes Good Yes No Bad So-so GoodPoor Decision Tree (bank loan)
10
10 Customer Data Name ____ Address ____ Years at address__ Co-applicant___ Job History Employer, Salary, Date Hired... Job History Employer, Salary, Date Hired... Loan Details Purpose Boat Loan Amount _____ Time _____ Data for Boat Loans Length: Engine: Cost New: Cost Used: Recommendation Lend $$$$ at ___ interest rate for ___ months, with ___ initial costs. Rules Frame-Based ES
11
11 ES Examples United AirlinesGADS: Gate Assignment American ExpressAuthorizer's Assistant StanfordMycin: Medicine/blood DECOrder Analysis + more Oil exploration Geological survey analysis IRS Audit selection Auto/Machine repair(GM:Charley) Diagnostic
12
12 ES Problem Suitability Narrow, well-defined domain Solutions require an expert Complex logical processing Handle missing, ill-structured data Need a cooperative expert
13
13 ES screens seen by user Rules and decision trees entered by designer Expert Forward and backward chaining by ES shell Knowledge engineer Knowledge database (for (k 0 (+ 1 k) ) exit when ( ?> k cluster-size) do (for (j 0 (+ 1 j )) exit when (= j k) do (connect unit cluster k output o -A to unit cluster j input i - A ))... ) Maintained by expert system shell Programmer Custom program in LISP ES Development ES Shells Guru Exsys Custom Programming LISP PROLOG
14
14 Limitations of ES Fragile systems Small environmental. changes can force revision. of all of the rules. Mistakes Who is responsible? Expert? Multiple experts? Knowledge engineer? Company that uses it? Vague rules Rules can be hard to define. Conflicting experts With multiple opinions, who is right? Can diverse methods be combined? Unforeseen events Events outside of domain can lead to nonsense decisions. Human experts adapt. Will human novice recognize a nonsense result?
15
15 AI Research Areas Computer Science Parallel Processing Symbolic Processing Neural Networks Robotics Applications Visual Perception Tactility Dexterity Locomotion & Navigation Natural Language Speech Recognition Language Translation Language Comprehension Cognitive Science Expert Systems Learning Systems Knowledge-Based Systems
16
16 Output Cells Sensory Input Cells Hidden Layer Some of the connections 3 -2 7 4 Input weights Incomplete pattern/missing inputs. Neural Network: Pattern recognition
17
17 Machine Vision Example The Department of Defense has funded Carnegie Mellon University to develop software that is used to automatically drive vehicles. One system (Ranger) is used in an army ambulance that can drive itself over rough terrain for up to 16 km. ALVINN is a separate road-following system that has driven vehicles at speeds over 110 kph for as far as 140 km.
18
18 Speech Recognition Look at the user’s voice command: Copy the red, file the blue, delete the yellow mark. Now, change the commas slightly. Copy the red file, the blue delete, the yellow mark. I saw the Grand Canyon flying to New York. Emergency Vehicles No Parking Any Time
19
19 Subjective Definitions temperature reference point e.g., average temperature coldhot Moving farther from the reference point increases the chance that the temperature is considered to be different (cold or hot). Subjective (fuzzy) Definitions
20
20 DSS, ES, and AI: Bank Example Decision Support SystemExpert SystemArtificial Intelligence NameLoan#LateAmount Brown25,000 51,250 Jones62,000 1 135 Smith83,000 32,435... Data Income Existing loans Credit report Model Lend in all but worst cases Monitor for late and missing payments. Output ES Rules What is the monthly income? 3,000 What are the total monthly payments on other loans? 450 How long have they had the current job? 5 years... Should grant the loan since there is only a 5% chance of default. Determine Rules loan 1 data: paid loan 2 data: 5 late loan 3 data: lost loan 4 data: 1 late Data/Training Cases Neural Network Weights Evaluate new data, make recommendation. Loan Officer
21
21 Decision Support SystemExpert SystemArtificial Intelligence Data a estimate sales K order setup cost h estimate holding cost Model Q* = sqrt ( 2ak / h ) Output time Q* Inventory Levels reorder points Choosing an Inventory System What is the cost of running out of inventory? 45,000 per day What are daily profits? 250,000 How many suppliers are there? 8 Can more suppliers be added in an emergency? no How close is the nearest supplier? 10 kilometres How reliable is this supplier? very... Best choice is to use Just-In-Time inventory system. Only a 2% chance of running out of inventory for more than 2 days.... Automatically Analyze site 1 data: JIT site 2 data: EOQ site 3 data: JIT site 4 data: hybrid Data/Training Cases Neural Network Weights Evaluate new data, make recommendation. DSS, ES and AI: Inventory Example
22
22 Vacation Resorts Software agent Resort Databases Locate & book trip. Software Agents Independent Networks/Communication Uses Search Negotiate Monitor
23
23 AI Questions What is intelligence? Creativity? Learning? Memory? Ability to handle unexpected events? More? Can machines ever think like humans? How do humans think? Do we really want them to think like us?
24
24 Appendix: Optimization Sample Problem You have three factories (Cheap, Intermediate, and Expensive). Output from the factories can be mixed and is measured in kilotons. You just received two orders to fill in the same time frame. How much production should come from each plant to fill the orders and minimize total costs?
25
25 Optimization Setup Variables: things you can control Output from each factory (c, i, e) given to each order (1, 2): Qc1, Qi1, Qe1, Qc2, Qi2, Qe2 Goal: objective function Minimize total cost: Cost = 75 Qc1 + 75 Qc2 + 100 Qi1 + 100 Qi2 + 120 Qe1 + 120 Qe2 Constraints: restrictions or limits Plant Limit C:Qc1 + Qc2 <= 100 Plant Limit I:Qi1 + Qi2 <= 100 Plant Limit E:Qe1 + Qe2 <= 100 Quantity Order 1:Qc1 + Qi1 + Qe1 = 100 Quantity Order 2:Qc2 + Qi2 + Qe2 = 150 Quality Order 1:2.0 Qc1 + 2.5 Qi1 + 3.0 Qe1 > 250(2.5 * 100) Quality Order 2:2.0 Qc2 + 2.5 Qi2 + 3.0 Qe2 > 225(1.5 * 150)
26
26 Spreadsheet Layout =C3*C2+D3*D2+E3*E2+F3*F2+G3*G2+H3*H2 =C4*C2+D4*D2 Put one variable in a column. Put objective function in a row. Enter formula for the objective value: sum of coefficient times variable. Put constraints in following rows with limit value. Enter each constraint formula: sum of coefficient times variable.
27
27 Excel Solver
28
28 Solution Plant output for order 1: C=28.555, I=41.327, E=30.118 Plant output for order 2: C=71.445, I=58.673, E=19.882 Total cost: 23,500 You can easily change the parameters and resolve the problem.
Similar presentations
© 2025 SlidePlayer.com. Inc.
All rights reserved.