Métodos de Apoio à Decisão

Some knowledge of Mathematical Programming.

Lectures introduce fundamental concepts, principles, and methods, supported by illustrative examples that contextualize the application of decision support techniques. Practical classes focus on structured problem solving, individually or in groups, promoting the development of analytical reasoning skills, selection of appropriate methods and critical interpretation of results.
Students engage in activities, including solving additional exercises, and preparing a group project involving decision-support methods, a written technical report and an oral presentation.

The course aims to equip students with the ability to model, analyse, and solve decision-support problems within Engineering and Industrial Management. Students are expected to understand and formulate deterministic and nondeterministic models, select appropriate solution methods, and assess underlying assumptions. They will develop skills in linear and nonlinear optimization, decision-making under uncertainty, utility theory, queueing models and decision trees. By the end of the course, students should also be able to interpret results, perform sensitivity analyses, and communicate technical conclusions in a rigorous and well-founded manner.

I-Deterministic models
1. Formulation and model development. Models based on linear programming. Sensitivity analysis. Integer programming-based models. 2. Optimization in networks: some models and algorithms. 3. Non-linear models. Formulation of problems. Optimization without restrictions on one or more variables. KKT conditions to nonlinear optimization with constraints. Programming: separable convex quadratic, non-convex.

II-Nondeterministic models
1. Decision theory. Probabilistic and non-probabilistic methods. Decision criteria in uncertainty. Value of information. Utility, indifference and risk. Risk premium. Decision criteria with risk. 2. Queueing models.

III-Decision trees
Decision nodes, states and alternatives. Selection, qualification and assessment of alternatives. Bayes analysis in estimating probabilities. Information value. Sensitivity analysis. Contingency plans.

Evaluation can be periodic or by final exam.
Periodic evaluation (to be carried out during classes):
· Test 1 (theoretical-practical, about deterministic models): quoted for 7.0 points;
· Test 2 (theoretical-practical, about nondeterministic models and decision trees): quoted for 7.0 points;
· Work (with written report and oral presentation): quoted for 6.0 points.
Requires that the sum of the two tests be greater than or equal to 7.0 points out of 14.
The student is approved if the sum of the work and both tests is greater than or equal to 9.5 points and meets the required minimum.

Evaluation by final exam:
Exam, theoretical-practical, quoted for 20 points, to perform during exam periods. The student is approved if he/she obtains a grade greater than or equal to 9.5 points.

NAO