Download as docx, pdf, or txt
Download as docx, pdf, or txt
You are on page 1of 6

Industrial automation :

Industrial automation is the set of technologies using


electronics, electrical engineering, mechanics,
telecommunications in order to design machines or automated
processes that can operate without human intervention.
Industrial automation can be found in several sectors such as
the food industry with packaging lines and special machines,
chemicals, the petroleum industry, electricity production plants,
the pharmaceutical industry, etc. industrial corresponds to
sequential automation and covers all the instrumentation and
control systems used to supervise or control a production chain.

the automation engineer must have knowledge of several


modules which are the basis of automation
and among this modules we have optimization.
what is optimization :
1- The theory of finding optimal points in a system
(maxima,minima)
2- A set of algorithms which can be applied to a mathematical
model of a system to find the :
a) Best design
-solving for the number of pipes in a piping network
that result in minimal pressure drop
-solving for the size of an engine block that results
in the best combination of speed and efficiency
( using some weighting on each of these metrics to
determine its importance )
b) lowest cost
-find the combination of truck routes to transport a set
amount of raw material between points that results in the lowest
total cost (i.e.,labor+truck maintenance+fuel)
c) highest efficiency
-Maximizing the efficiency of a power production system by
finding the right combination of inputs (flow rates
,temperatures , pressures,...etc)
The role of modeling in optimization :
1) Optimization involves “ testing out “ many different
combinations of inputs to determine what the resulting
output will be under each of those circumstances .
2) With many inputs, it is usually not paratical to just guess
and check on an actual system
3) Optimization requires a mathematical model of the system,
which is just the math that relates the inputs to the outputs
4) By using a model of the system to iterate, there is really no
limit to the number of combinations you can try
5) Optimization algorithms are really much more sophisticated
than just guessing and checking a range of
combinations .optimization theory has algorithms that are
custom-made to find the optimal solution with minimal
computation by exploiting knowledge of the model
The anatomy of an optimization problem :

Types of optimization problems :


1) Linear program (LP) : objective function and all constraints
are linear w.r.t. the decision variables
2) Quadratic program (QP) : objective function is quadratic
w.r.t. the decision variables and all constraints are linear
w.r.t. the decision variables
3) Nonlinear program (NLP) : objective function is nonlinear
w.r.t. the decision variables and constraints can be either
linear or nonlinear functions of the decision variables
4) Mixed-integer program : some of the constraints are
restricted to discrete values of decision variables (e.g.,0or
1, on or off , only integers : 0,1,2,3,4…
 Mixed-integer linear program (MILP)
 Mixed-integer nonlinear program (MINLP)
How to slove an optimization problem :

 develop a model od your system


 consider the model complexity
 consider the model form(e.g,linear,quadratic,nonliner,mixed-
intger,etc.->these are listed in order of increasing
complexity,generally speaking)
 Determine your objective function (i.e,what are you trying to minimize or
maximize)
 this is usually a combination of additive terms,all put on the same
basis (e.g,profite=sum of revenue-sum of costs)
 by convention ,most solver will require your problem to be a
minimization problem.for exmple,if you want to maximize
profit,you actually need to formulate your problem so that-profit is
minimized.
 Determine your constraints(e.g.,a flow rate >=0,Temperature<=some
explosive limit,etc.)
 Put the math representing your model into the form required by the
solver
 this typically takes some very careful manipulation
 many solvers require putting the problem into matrix form
 Setting up the problem in the right format is most of the battle
the anatomy of an Optimization Problem

min f(x)
x

Objective
fonction

Decision variables
contrains on décision
(in a vector)
s.t LB ≤ x ≤ UB variables

g(x)=0
Qualité contraints

h(x) ≤ 0
x1

[]
inequlity construits
x2
x= x3
x4

 development

You might also like