Innovative computational intelligence ai techniques - Ahmed Yousry

Innovative Computational Intelligence: AI
Techniques
Ahmed Yousry Abdelsatar
Assist Lecturer at BFCAI
PHD Student
ComputerScience Department.
Presented by:

Agenda
❑What is optimization.
❑Cuckoo search algorithm.
❑Shuffled frog leaping algorithm.
❑Harmony Search Algorithm.
❑Gravitational Search Algorithm.
❑Fruit Fly Optimization Algorithm
❑Intelligent Water Drops Algorithm

What is optimization....?
• It’s a procedure to make a system or design more effective, especially
involving the mathematical techniques.
➢To minimize the cost of production .
➢to maximize the efficiency of production.
• It’s a technique to:
➢ Find Best Solution
➢ Minimal Cost
➢ Minimal Error
➢ Maximal Profit
➢ Maximal Utility

Methods of Optimization
❑Mathematical optimization
1) Linear Programming
2) Dynamic Programming
❑Evolutionary algorithms-
That mimic the metaphor of natural biological evolution and the social
behavior of species.

Cuckoo search algorithm
• A method of global optimization based on the behavior of cuckoos
was proposed by Yang & Deb (2009).
• The original “cuckoo search (CS) algorithm” is based on the idea of the
following :-
➢How cuckoos lay their eggs in the host nests.
➢How, if not detected and destroyed, the eggs are hatched to chicks by the hosts.
➢How a search algorithm based on such a scheme can be used to find the global
optimum of a function.

Behavior of Cuckoo breeding
❑The CS was inspired by the parasitism ‫تطفل‬ of some
cuckoo species by laying their eggs in the nests of host
birds.
❑Some cuckoos have evolved in such a way that female
parasitic cuckoos can imitate the colors and patterns of
the eggs of a few chosen host species.
❑This reduces the probability of the eggs being
abandoned ‫تهجر‬ and, therefore, increases their
reproductivity .

Behavior of Cuckoo breeding
❑If host birds discover the eggs are not their own, they will
either throw them away or simply abandon their nests
and build new ones.
❑Parasitic cuckoos often choose a nest where the host bird
just laid its own eggs.
❑In general, the cuckoo eggs hatch slightly earlier than
their host eggs.

Behavior of Cuckoo breeding cont…
❑Once the first cuckoo chick is hatched, his first instinct
action is to evict ‫طرد‬ the host eggs by blindly propelling the
eggs out of the nest.
❑This action results in increasing the cuckoo chick’s share of
food provided by its host bird .
❑Moreover, studies show that a cuckoo chick can imitate the
call of host chicks to gain access to more feeding
opportunity.

Characteristics of Cuckoo search
❑Each egg in a nest represents a solution, and a cuckoo
egg represents a new solution.
❑The aim is to employ the new and potentially better
solutions (cuckoos) to replace not-so-good solutions in
the nests.
❑In the simplest form, each nest has one egg.
❑The algorithm can be extended to more complicated
cases in which each nest has multiple eggs representing
a set of solutions.

Characteristics of Cuckoo search (cont)
The CS is based on three idealized rules:
1. Each cuckoo lays one egg at a time, and dumps it in a
randomly chosen nest
2. The best nests with high quality of eggs (solutions) will
carry over to the next generations.
3. The number of available host nests is fixed, and a host can
discover an strange egg with probability p ϵ [0,1] .
❑In this case, the host bird can either throw the egg away or
abandon the nest to build a completely new nest in a new
location.

Lѐvy Flights
❑In nature, animals search for food in a random or quasi-
random manner.
❑Generally, the path of an animal is effectively a random
walk because the next move is based on both the current
location/state and the transition probability to the next
location.
❑The chosen direction implicitly depends on a
probability, which can be modeled mathematically.
❑A Lévy flight is a random walk in which the step-
lengths are distributed according to the probability
distribution.
❑After a large number of steps, the distance from the
origin of the random walk tends to a stable distribution.

Cuckoo searchAlgorithm
Step1. Generate initial population of n host nests.
(𝑎𝑖,𝑟𝑖) : a candidate for optimal parameters.

Step2. Lay the egg (𝑎𝑘′,𝑏𝑘′) in the k nest.
K nest is randomly selected.
Cuckoo’s egg is very similar to host egg.
Where
𝑎𝑘′=𝑎𝑘+𝑅𝑎𝑛𝑑𝑜𝑚𝑤𝑎𝑙𝑘(Lѐvy flight)𝑎𝑘
𝑟𝑘′=𝑟𝑘+𝑅𝑎𝑛𝑑𝑜𝑚𝑤𝑎𝑙𝑘(Lѐvy flight)𝑟𝑘
Cuckoo searchAlgorithm cont…

Step3. Compare the fitness of cuckoo’s egg with the
fitness of the host egg.
Root Mean Square Error (RMSE)

Cuckoo searchAlgorithmCont.…
• Step4. If the fitness of cuckoo’s egg is better than host egg, replace
the egg in nest k by cuckoo’s egg.

Step5. If host bird notice it, the nest is abandoned and new one is built. (p <0.25) (to
avoid local optimization)
❑ Iterate steps 2 to 5 until termination criterion satisfied

Application of the CS Algorithm
❑Engineering optimization problems
❑NP hard combinatorial optimization problems
❑Data fusion in wireless sensor networks
❑Nano electronic technology based operation-amplifier (OP-AMP).
❑Train neural network.
❑Manufacturing scheduling
❑Nurse scheduling problem

Shuffled frog leaping algorithm
❑Shuffled frog leaping algorithm (SFLA) was recently proposed in (Eusuff and
Lansey; Eusuff et al.; Eusuff 2004) for solving problems with discrete
decision variables.
❑Inspired by natural memetics ‫,التطور‬ SFLA is a population-based cooperative
search metaphor combining the benefits of the genetic-based memetic
algorithm (MA) and the social behavior based particle swarm
optimization(PSO).

What is shuffled frog leaping algorithm...?
❑The SFLA is a method which is based on
observing, imitating, and modelling the behavior
of a group of frogs when searching for the location
that has the maximum amount of available food .

Process of SFLA
❑Population consists of a set of frogs Partitioned into subsets referred to
as memeplexes.
❑Each memeplexes performing a local search After a defined number of
evolution steps, ideas are passed among memeplexes in a shuffling
process.
❑The local search and the shuffling processes continue until defined
convergence criteria are satisfied

Analytical Process
▪ Population of P frogs is created randomly.
▪ A frog i is represented as xi (xi1, xi2,., Xi), Sorted in a descending order
according to their fitness.
▪ Population is divided into m memeplexes, each containing n frogs, P=m × n
▪ Frogs with the best and the worst fitness are identified as xb and xw.
▪ Change in frog position (Di) = rand( )× (xb-xw).
▪ Previous position Xw , New position (Xb)= Xw + Di.
▪ If no improvement becomes possible in this case, then a new solution is
randomly generated to replace that frog.
▪ The calculations then continue for a specific number of iterations.

Frog CallingAlgorithm
❑Factor 1: A frog will first check that if there is any calling frog in its own
territory range, and then it will confirm that if the total number of calling
frogs existing in the paddy field ‫حقل‬ is still within an acceptable range. Once
it is done with these, it will decide to produce calls or not.
❑Factor 2: Number of competing frogs. A frog will evaluate its surroundings
and compare itself with other calling frogs according to some criteria. If the
probability for the frog to win is high, it will begin to call anyway.
❑Factor 3: Body size. Once the weak calling frog detects its current condition,
it will adopt sleep strategy to avoid competition.

Application of SFLA
❑Ac-dc optimal power flow.
❑Scheduling of construction projects.
❑Computer-aided design activities.
❑Water distribution network design.

Advantages
❑SFLA has been used as appropriate tools to obtain the best solutions
with the least total time and cost by evaluating unlimited possible
options.
❑Implementation of evolutionary algorithms in various field because of
their reliability and simple implementation

Harmony search Algorithm
❑Harmony search (HS) is a population based metaheuristics
algorithm inspired from the musical process of searching for a
perfect state of harmony.
❑HS has been proposed by Geem et al. in (2001)
❑The pitch of each musical instrument determines the
aesthetic ‫جمالى‬ quality, just as the fitness function value
determines the quality of the decision variables.
❑In the music improvisation process, all players sound pitches
within possible range together to make one harmony.

Harmony search Algorithm cont…
❑If all pitches make a good harmony, each player stores in
his memory that experience and the possibility of making,
a good harmony is increased next time.
❑The same thing in optimization, the initial solution is
generated randomly from decision variables within the
possible range.
❑If the objective function values of these decision variables
is good to make a promising solution, then the possibility
to make a good solution is increased next time.

Initialization of harmony memory
❑The initial population HM contains of HMS vectors is generated
randomly, where
xi = xij , i = 1, …,HMS and j = 1, …, n.
❑ The HM matrix is filled with HMS vectors as follows:

Improvisation of new harmony vectors
❑Harmony memory considering (HMC) rule.
❑For this rule, a new random number r1 is generated within the range [0,1].
❑If r1 < HMCR, where HMCR is the harmony memory consideration rate,
then the first decision variable in the new vector xij {new} is chosen
randomly from the values in the current HM as follows:

Improvisation of new harmony vectors (cont..)
➢Pitch adjusting rate (PAR).
❑ A new random number r2 is generated within the range [0 1].
❑ If r2 < PAR, where PAR is a pitch adjustment rate, then the pitch
adjustment decision variable is calculated as follows:
❑ where BW is a bandwidth factor, which is used to control the local
search around the selected decision variable in the new vector.

Improvisation of new harmony vectors (cont)
➢Random initialization rule
❑If the condition r1 < HMCR fails, the new first decision variable in the new
vector
❑x ij {new} is generated randomly as follows:
• where l, u is the lower and upper bound for the given problem.

Harmony memory updating
❑After the harmony vector x{new} is generated, it will replace the worst
harmony vector x{worst} in the harmony memory if its objective
function value is better than the objective function value of the worst
harmony vector.

Application of the HS Algorithm
❑Engineering optimization problems.
❑NP hard combinatorial optimization problems.
❑Data fusion in wireless sensor networks.
❑Nano electronic technology based operation-amplifier (OP-AMP).
❑Train neural network.
❑Manufacturing scheduling.
❑Nurse scheduling problem.

Gravitational search algorithm
❑Gravitational search algorithm (GSA) is a population
search algorithm proposed by Rashedi et al. in 2009.
❑The GSA is based on the low of gravity and mass
interactions.
❑The solutions in the GSA population are called
agents, these agents interact with each other through
the gravity force.
❑The performance of each agent in the population is
measured by its mass.

❑Each agent is considered as object and all objects
move towards other objects with heavier mass due to
the gravity force.
❑This step represents a global movements (exploration
step) of the object, while the agent with a heavy mass
moves slowly, which represents the exploitation step
of the algorithm.
❑The best solution is the solution with the heavier
mass.

Gravitational search algorithm cont…
❑The gravitational constant G at iteration t is computed as
follows.
❑Where G0 and 𝛼 are initialized in the beginning of the
search, and their values will be reduced during the search.
T is the total number of iterations.

The gravity low
❑The objects masses are obeying the low of gravity as following
❑Equation represents the Newton law of gravity, where
F is a magnitude of the gravitational force
G is gravitational constant
M1 is the mass of the first object
M2 is the mass of the second object
R is the distance between the two objects M1, M2.

The gravity low cont…
❑According to the Newton's second low, when a
force F is applied to an object, the object moves
with acceleration a depending on the applied
force and the object mass M as shown in
Equation.

Acceleration of agents
❑There are three kind of masses
➢Active gravitational mass Ma
➢Passive gravitational mass Mp
➢Inertial mass Mi.
❑The gravitational force Fij that acts on mass i by mass j is defined
by:
Where Maj, Mpi are the active and passive masses of objects j, i, respectively.

Acceleration of agents (Cont.)
• The acceleration of object (agent) i is computed
as follows.
Where Mii is inertia mass of agent i.

Agent velocity and positions
• During the search, the agents update their
velocities and positions as shown in Equations
respectively.

• The main steps of the GSA can be summarized as follows.
➢ Step 1. The algorithm starts by setting the initial values of
gravitational constant G0, 𝛼, 𝜀 and the iteration counter t.
➢ Step 2. The initial population is generated randomly and consists
of N agents, the position of each agent is defined by:

➢Step 3. The following steps are repeated until termination
criteria satisfied
▪ Step 3.1. All agents in the population are evaluated and the best, worst
agents are assigned.
▪ Step 3.2. The gravitational constant is updated.
▪ Step 3.3. When agent j acts on agent i with force, at a specific time
(t) the force is calculated as following:
Where Maj is the active gravitational mass of agent j, Mpi is the passive
gravitational mass of agent i, G(t) is gravitational constant at time t

Gravitational search algorithm (Cont.)
▪ Step 3.4. At iteration t, calculate the total force acting on agent i as
following:
Where Kbest is the set of first K agents with the best fitness value and biggest mass.
▪ Step 3.5. Calculate the inertial mass as following:

Gravitational search algorithm (Cont.)
▪ Step 3.6. The acceleration of agent i is calculated as following:
▪ Step 3.7. The velocity and the position of agent i.
▪ Step 3.8. The iteration counter is increased until termination
criteria satisfied.
➢Step 4. The best optimal solution is produced.

Fruit Fly OptimizationAlgorithm
• Fruit flies are small flies and usually with red eyes. They are especially
attracted to ripened foods in the kitchen. They can even smell food source
from 40 km away(Pan 2012).
• In addition, the number of the fruit fly’s eye (i.e., compound eye) are huge in
which contains 760 unit eyes (Chapman 2013).
• Based on those characteristics, the fruit fly can exploit an extra ordinarily
wide range of food sources

food finding process of fruit fly
• It smells the food source by osphresis organ ‫الشم‬ ‫,حاسة‬ and flies towards that
location.
• After it gets close to the food location, the sensitive vision is also used for
finding food and other fruit flies’ flocking location.
• finally, it flies towards that direction.

Fundamentals of Fruit Fly Optimization
Algorithm
1. Initialization phase: The fruit flies are randomly distributed in the search
space(Init X_axis and Init Y_axis)
Where the term ‘‘RandomValue’’ is a random vector that were sampled from a
uniform distribution.

Cont…
2. Path construction phase: The distance and smell concentration value of
each fruit fly can be defined via
Where Disti is the distance between the ith individual and the food location, and Si is the smell
concentration judgment value which is the reciprocal ‫العكسية‬ ‫القيمه‬of distance.
3. Fitness function calculation phase. It can be defined via
Where Smelli is the smell concentration of the individual fruit fly, bestSmell and bestIndex represent the
largest elements and its indices along different dimensions of smell vectors, and max(Smelli) is the
maximal smell concentration among the fruit flies.

Cont…
4. Movement phase: The fruit fly keeps the best smell concentration value and
will use vision to fly towards that location via Eqs

Implementing FFOA can be summarized as
follows
Step 1: Initialize the optimization problem and algorithm parameters.
Step 2: Repeat till stopping criteria met. First, randomly select a location via
distance and smell concentration judgment value. Second, calculate its fitness
function Function(Si). Third, find out the fruit fly with maximal smell
concentration among the fruit fly swarm. Fourth, rank the solutions and move to
the best solution.
Step 3: Post process and visualize results.

The main advantages of FFOA
❑Include simple computational process.
❑Ease understanding.
❑Easy implementation.

Intelligent waterdrop algorithm
❑Intelligent Water Drops algorithm, or the IWD algorithm, is a Swarm
based nature-inspired optimization algorithm.
❑This algorithm contains a few essential elements of natural water
drops and actions and reactions that occur between river's bed and the
water drops that flow within.
❑The IWD algorithm can be used for optimization.

Natural water drops
❑In nature, flowing water drops are observed mostly in rivers, which form
huge moving swarms.
❑ The paths that a natural river follows have been created by a swarm of
water drops.
❑ For a swarm of water drops, the river in which they flow is the part of the
environment that has been dramatically changed by the swarm and will
also be changed in the future.
❑We are using the concept of the water path in the river. How they prefers a
path with less soil than a path with more soil.
❑The water drop prefers an easier path to a harder path when it has to
choose between several branches.

Implementation of IWD algorithms
❑The IWD algorithm employs a number of IWDs to find the optimal solutions
to a given problem.
❑The problem is represented by a graph (N, E) with the node set N and edge
set E.
❑This graph is the environment for the IWDs and the IWDs flow on the edges
of the graph.
❑Each IWD begins constructing its solution gradually by traveling between
the nodes of the graph along the edges until the IWD finally completes its
solution denoted by T IWD
❑Each solution T IWD is represented by the edges that the IWD has visited.

IWD vs ANT COLONY
❑Every ant in an Ant Colony Optimization (ACO) algorithm
deposits pheromones on each edge it visits. In contrast, an
IWD changes the amount of soil on edges.
❑In the ACO algorithm, an ant cannot remove pheromones
from an edge whereas in the IWD algorithm, an IWD can
both remove and add soil to an edge.
❑Besides, the IWDs may gain different velocities throughout
an iteration of the IWD algorithm whereas in ACO
algorithms the velocities of the ants are irrelevant.

Performance of IWD
❑
▪
▪
▪
▪
❑

Innovative computational intelligence ai techniques - Ahmed Yousry

Recomendados

Recomendados

Mais conteúdo relacionado

Semelhante a Innovative computational intelligence ai techniques - Ahmed Yousry

Semelhante a Innovative computational intelligence ai techniques - Ahmed Yousry (20)

Último

Último (20)

Innovative computational intelligence ai techniques - Ahmed Yousry