Bacterial Foraging Optimization - PowerPoint Presentation

Slide1

Bacterial Foraging Optimization

Presented by: Adia KhalidPhD (Scholar) Home Energy Optimization in Smart GridSupervised by: Dr. Nadeem Javaid

1Slide2

Introduction

Bacterial Foraging AlgorithmHome Energy Management Agenda

2Slide3

Natural selection method

eliminate animals with poor “foraging strategies” Foraging strategiesmethods for locating, handling, and ingesting food favor those animals that have successful foraging strategies

obtain enough food to enable them to reproduceafter many generations poor foraging strategies are:

either eliminated orshaped into good ones i.e.. Redesignedsuch evolutionary principles have led scientists in the field of “foraging theory” to hypothesize

it is appropriate to model the activity of foraging as an optimization process

Introduction (1\4)

3

Passino

, K. M. (2002). Biomimicry of bacterial foraging for distributed optimization and control. 

IEEE control systems

, 

22

(3), 52-67.Slide4

Introduction (2\4)

A foraging animal main focus: maximize the energy obtained per unit time spent foragingIn the face of constraints presented by its own physiology e.g., sensing and cognitive capabilities

environment e.g., density of prey, risks from predators, physical characteristics of the search area

Evolution has balanced these constraints sometimes referred to asoptimal foraging policy

such terminology is especially justified in case

where the models and policies have been ecologically validated

4

Foraging focus

Energy Maximization

Constraints

Physiology

Environment

Passino

, K. M. (2002). Biomimicry of bacterial foraging for distributed optimization and control. 

IEEE control systems

, 

22

(3), 52-67.Slide5

Introduction (3\4)

Optimal Foraging:Optimal foraging theory formulates the foraging problem as an optimization problem.Optimization models valid for social foragingwhere groups of animals cooperatively forage

Here, we explain the biology and physics underlying the chemotactic (foraging) behavior of E.coli bacteria for optimization named as Bacterial Foraging Optimization

Individual bacterium also communicates with others by sending signalsDuring foraging of a real bacteria two basic operations

Swim

Tumble

performed by a bacterium at the time of foraging by

a set of tensile flagella

5

http://wikis.swarthmore.edu/mathbio/index.php?title=File:Optimal_Foraging_Theory.jpg&limit=50

Passino

, K. M. (2002). Biomimicry of bacterial foraging for distributed optimization and control. 

IEEE control systems

, 

22

(3), 52-67.Slide6

Introduction (4\4)

E.coli bacteria: the ones that are living in our intestines

StructureDiameter: 1μmLength: 2μm

Flagellum:Chemotaxes movement in the presence of

chemical attractants and repellants

Counterclockwise:

Swim

Clockwise:

Tumble

Bactria

chemotactic

Movement

6Slide7

Bacterial Foraging Algorithm (1\2)

7

Chemotaxis:

When a bacterium meets a favorable environment (rich in nutrients and noxious free),

it will continuous swimming in the same direction

,

if nutrient increase in the direction of swim

when it meets an unfavorable environment,

it will tumble, i.e., change the direction of swim

Reproduction:

After

calculating fitness value for each bacteria,

reproduction allows the bacteria to survive and reproduce

Swarming:

E. Coli bacterium has a specific sensing actuation and decision making mechanism

On each move

It releases attracts to signal other bacteria to swarm towards its

Elimination and dispersal:

The

chemotaxis provides a basis for local search and reproduction speeds the convergence

To avoid trap bacteria in local minimum

Elimination-Dispersal is done

tumble

run

run

tumble

run

tumbleSlide8

The algorithm models bacterial

Population: {i ϵ pop, where pop have 1:Np elements}Chemotaxis: {for j=1:Nc elements}

Swarming: {for s=1:Ns elements}Reproduction: {for k=1:N

r elements}Elimination and Dispersal Steps: {for l=1:Ne step}

Another parameter

C: Step Size for the dimension

8

Bacterial Foraging Algorithm (2\2)Slide9

BFA Mapping on Home Energy Management (HEM)

Target: We have to optimize the energy usage By scheduling the home appliancesPass the on-peak hour load on off-peak hour

Home Energy Management (1/10)

9Slide10

BFA Mapping on Home Energy Management (HEM)

Optimal Solution get form optimal search spaceSearch space Population of group of bacteria's

Home Energy Management (2/10)

10

BFA Parameters

HEM Parameters

Values

Population

Possible solution

30

Bactrians with in a swarm

Appliances

9

Bacterium

status i.e. dead or alive

Appliance’s ON

or OFF status

1 or 0

Elimination steps

Schedule

24 hour

Fitness Level i.e. min( )

Min (Cost)

vary

Cost depend on the power rate of an appliance and price signals Slide11

Peak Load Reduction

Demand Responseencourage the user to make changes in their demand according to the price signals Implementation is possibly implemented using some energy measuring consumption measuring meters:Bulk meter just save information with a single bulk

Smart meter record electricity usage information frequently

[*]. Waterloo North Hydro, https://www.wnhydro.com/en/your-home/time-of-use-rates.asp. Last visited: 20 May 2016.

Home Energy Management (3/10)

11

Pricing schemes:

Flat Rates 

-

same rate during a given period of time

E.g., 30-day bill cycle

Tiered Rates

 - charge a different price based on blocks of usage

e.g., first 500 kWh vs. next 500 kWh for 30-day billing cycle

Time-based rate includes-

Dynamic Rate and

Time of Use (TOU)

Dynamic Rate

[*]

Real Time Price (RTP)

Critical Peak Pricing (CPP)

Curtailable/Interruptible (C/I)

Variable Peak Pricing (VPP)

Critical Peak Rebate (CPR)Slide12

Pricing Scheme

[*]TOU: Prices set for the on-peak and off-peak hours, where hours divided into blocks and price for a particular block remain fixedRTP: Rates tariff based on the hourly bases usage of electricity

Utilities regulates RTP in two parts Base bill calculated on the bases of define tariff for particular customer depend on

customer baseline load (CBL)2) Hourly prices apply according to the customer usage that is a difference between actual and CBL

……………….. (1)

The C/I Option:

Specifies conditions under which disturbance in service may occur

A customer gives right to the Local Distribution Company (LDC) during the contract

not the obligation to disturb the services

In such situation LDC pays incentive to the customers through bill reduction

[*]. Waterloo North Hydro, https://www.wnhydro.com/en/your-home/time-of-use-rates.asp. Last visited: 20 May 2016.

Home Energy Management (4/10)

12Slide13

Pricing Scheme

[*]VPP: hybrid TOU and RTP

=

+

……………. (2)

……………...(2a)

……………...(2b)

CPP:

Critical events may call during the specific period

utilities observe high market price rate or during the power system emergency conditions

Usually occur in hot summer weekdays

Allow only 15 time per year

CPR:

During the critical event utilities increases the price for the specific time duration

refunded it to customer

when utilities observe any reduction in consumption

[*]. Waterloo North Hydro, https://www.wnhydro.com/en/your-home/time-of-use-rates.asp. Last visited: 20 May 2016.

Home Energy Management (5/10)

13Slide14

BFA

Matlab Code

Home Energy Management (6/10)

14

Bacterium

\ Appliances

Fitness Evolution

Vacuum

Cleaner

Water Heater

Water

Pump

Dish Washer

Refrigerator

AC

Oven

Washing Machine

Cloth Dryer

J

last

1

0

0

1

0

1

1

1

1

101

0

1

1

1

1

0

0

0

0

100

1

1

1

1

1

0

0

1

1

99

1

0

1

1

0

0

1

1

0

100

for

i=1:NP

%

for

j=1:D-1

if

rand(1)>0.6

X=1;

else

X=0;

end

J(

i

)=sum(100*(x(k,j+1)-x(

i,j

)^2)^2+(x(

i,j

)-1)^2);

% initial fitness calculation

end

endSlide15

BFA

Matlab Code

Home Energy Management (7/10)

15

Bacterium

\ Appliances

Fitness Evolution

Vacuum

Cleaner

Water Heater

Water

Pump

Dish Washer

Refrigerator

AC

Oven

Washing Machine

Cloth Dryer

J

0

1

0

0

0

1

1

0

0

100

1

1

1

0

0

1

1

1

0

110

1

1

0

1

1

1

0

1

0

106

0

0

1

0

1

0

1

0

0

105

for

l=1:Ne

% 24 l=elimination of dispersal step

for

k=1:Nr

% 4 k=reproduction

for

j=1:Nc

% 3 Chemotaxis Loop %

for

i

=1:Np

% 4 Take a chemotaxis Step

del=(rand(1,D)-0.5)*2;

x(

i

,:)=x(

i

,:)+(C/

sqrt

(del*del'))*del;

%

C= 0.4

%Direction of Tumble i.e. new position of bacterium

for

d=1:D-1

J(i)=sum(100*(x(i,d+1)-x(i,d)^2)^2+(x(i,d)-1)^2);

%Fitness Evalutin

end

Slide16

BFA

Matlab Code

Home Energy Management (8/10)

16

Bacterium

\ Appliances

Fitness Evolution

Vacuum

Cleaner

Water Heater

Water

Pump

Dish Washer

Refrigerator

AC

Oven

Washing Machine

Cloth Dryer

J

last

0

0

0

1

0

0

1

1

1

100

1

1

0

0

0

1

1

0

1

110

1

1

0

1

1

1

0

1

0

106

0

0

1

0

1

0

1

0

0

105

for

m=1:Ns

% 2 swimming Step

if

J(

i

)<

Jlast

(

i

)

%Elimination and Dispersal Check

Jlast

(

i

)=J(

i

);

x(

i

,:)=x(

i

,:)+C*(del/

sqrt

(del*del'));

%Direction of Tumble i.e. new position of bacterium

for

d=1:D-1

J(i)=sum(100*(x(i,d+1)-x(i,d)^2)^2+(x(i,d)-1)^2);

%Fitness Function

end

else

Updated the bacterium\ appliances status

del=(rand(1,D)-0.5)*2;

%Vector with random Direction

x(

i

,:)=x(

i

,:)+C*(del/

sqrt

(del*del'));

for

d=1:D-1

J(i)=sum(100*(x(i,d+1)-x(i,d)^2)^2+(x(i,d)-1)^2);

end

end

end

% end swimming Step

end

% end of bacterial populationSlide17

BFA

Matlab Code

Home Energy Management (9/10)

17

Bacterium

\ Appliances

Cost

Vacuum

Cleaner

Water Heater

Water

Pump

Dish Washer

Refrigerator

AC

Oven

Washing Machine

Cloth Dryer

Cost_B

(cent)

0

0

0

1

0

0

1

1

1

48

1

1

0

0

0

1

1

0

1

30

1

1

0

1

1

1

0

1

0

45

0

0

1

0

1

0

1

0

0

50

for

i

=1:Np

%% Check the Health

Cost_B

(

i

)= min(Cost);

%

power_Rate

*

Electricity_Price

end

end

% end of reproduction step

Minimum cost will be selected Slide18

BFA

Matlab Code

Home Energy Management (10/10)

18

Bacterium

\ Appliances

Fitness Evolution

Vacuum

Cleaner

Water Heater

Water

Pump

Dish Washer

Refrigerator

AC

Oven

Washing Machine

Cloth Dryer

J

1

1

0

1

0

1

1

1

0

103

1

0

0

1

0

0

1

0

1

106

0

1

0

1

0

1

0

1

1

100

0

1

1

1

1

0

0

0

0

109

%% random elimination dispersion

for

j=1:Np

for

i

=1:D

if

rand(1)>=Ped

x(

j,i

)=1;

else

x(

j,i

)=0;

end

for

d=1:D-1

J(i)=sum(100*(x(i,d+1)-x(i,d)^2)^2+(x(i,d)-1)^2);

end

end

end

endSlide19

19