ZEIT4700 – S1, 2015 Mathematical Modeling and Optimization School of Engineering and Information Technology
Optimization - basics Maximization or minimization of given objective function(s), possibly subject to constraints, in a given search space Minimize f1(x),..., fk(x) (objectives) Subject to gj(x) < 0, i = 1,...,m (inequality constraints) hj(x) = 0, j = 1,..., p (equality constraints) Xmin1 ≤ x1 ≤ Xmax1 (variable / search space) Xmin2 ≤ x2 ≤ Xmax2 (discrete/continuous/mixed).
Optimization - basics Maximization or minimization of an objective function, possibly subject to constraints x F(x) Local minimum Global Minimum (unconstrained) Constraint 2 (active) Constraint 1 Global Minimum (constrained)
Optimization - basics x1 x2 f1 f2 Variable spaceObjective space Linear / Non-linear / “Black-box”
Some considerations while formulating the problem Objective function(s) -- Should be conflicting if more than 1 (else one or more of them may become redundant). Variables – Choose as few as possible that could completely define the problem. Constraints – do not over-constrain the problem. Avoid equality constraints where you can (consider variable substitution / tolerance limits). f2 f1
Example Design a cylindrical can with minimum surface area, which can hold at least 300cc liquid.
Classical optimization techniques Region elimination (one variable) Gradient based Linear Programming Quadratic programming Simplex Drawbacks 1.Assumptions on continuity/ derivability 2.Limitation on variables 3.In general find Local optimum only 4.Constraint handling 5.Multiple objectives Newton’s Method (Image source : Nelder Mead simplex method (Image source :
Optimization – types / classification Single-objective / multi-objective Unimodal / multi-modal Single / multi - variable Discrete / continuous / mixed variables Constrained / unconstrained Deterministic / Robust Single / multi-disciplinary
Optimization - methods Classical Region elimination (one variable) Gradient based Linear Programming Quadratic programming Simplex Heuristic / metaheuristics Evolutionary Algorithms Simulated Annealing Ant Colony Optimization Particle Swarm Optimization.
Project 1 Nature optimizes both living and nonliving objects. Identify an object that has been optimized; Develop the mathematical formulation of what has been minimized/maximised and present results to justify why it has taken the form. (Due April 09, 2015)
Resources Course material and suggested reading can be accessed at Hemant/design-2.html Hemant/design-2.html