what is a metaheuristic

7/31/2019 What is a Metaheuristic

1/17

What is a Metaheuristic?

Spring 2009

Instructor: Dr. Masoud Yaghini


2/17

Optimization Problems


3/17


Optimization problems

The many decision-making problems can beoften expressed as an constrainedoptimization problem with some decisionvariables that are restricted by a set of

.

Types of constrained optimization problems:

Combinatorial problems: When the decisionvariables are discrete

Continuous problems: When the decisionvariables are continuous

Mixed problems


4/17


Combinatorial Problems Examples of real-world combinatorial

optimization problems include:

Assembly-line balancing problems

Vehicle routing and scheduling problems

Facility layout design problems

Job sequencing and machine scheduling problems

Manpower planning problems

Production planning and distribution

Etc.


5/17


Combinatorial Optimization Combinatorial optimization problems are often

easy to state but very difficult to solve.

Many of the problems arising in applicationsare NP-hard, that is, it is strongly believed that

polynomially bounded computation time.


6/17


Combinatorial Optimization Two classes of algorithms are available for the

solution of combinatorial optimizationproblems:

Exact algorithms


7/17


Combinatorial Optimization Exact algorithms are guaranteed to find the

optimal solution and to prove its optimality forevery finite size instance of a combinatorialoptimization problem within an instance-

.

In the case of NP-hard problems, in the worstcase, exponential time to find the optimum.

For most NP-hard problems the performanceof exact algorithms is not satisfactory.


8/17


Combinatorial Optimization If optimal solutions cannot be efficiently

obtained in practice, the only possibility is totrade optimality for efficiency.

Approximate algorithms, often also called,

to obtain good, that is, near-optimal solutionsat relatively low computational cost withoutbeing able to guarantee the optimality of

solutions.


9/17

Metaheuristics


10/17


Metaheuristics A disadvantage of heuristic methods is that

they:

either generate only a very limited number ofdifferent solutions, or

h r li l l im whi h i h

case for iterative improvement methods.

Metaheuristics have been proposed which tryto bypass these problems.

Metaheuristics apply to solve the problemsknown as of difficult optimization

Available from the 1980s


11/17


Metaheuristics Definition:

A metaheuristic is a set of algorithmic conceptsthat can be used to define heuristic methodsapplicable to a wide set of different problems.

A metaheuristic can be seen as a eneral-

purpose heuristic method toward promisingregions of the search space containing high-qualitysolutions.

A metaheuristic is a general algorithmic frameworkwhich can be applied to different optimizationproblems with relatively few modifications to makethem adapted to a specific problem.


12/17


Capability of Metaheuristics

Metaheuristics have capability to beextracted from a local minimum


13/17


Metaheuristics The metaheuristics are from now on regularly

employed in all the sectors of engineering,

Examples of metaheuristics algorithms: The evolutionary algorithms

The tabu search method

The ant colony optimization The simulated annealing method

Etc.


14/17


Metaheuristics


15/17

References


16/17


References J. Dreo A. Petrowski, P. Siarry E. Taillard,

Metaheuristics for Hard Optimization, Springer-

Verlag, 2006. R.J. Moraga, G.W. DePuy, G.E. Whitehouse,

Metaheuristics: A Solution Methodology forp m za on ro ems, an oo o n us r a an

Systems Engineering, A.B. Badiru (Ed.), 2006.


17/17

The End

what is a metaheuristic

Documents