Assigning different upper and lower bounds to function input variables when calling the function during optimization

I am using the PSO to optimize the RMSE of the computed and real energy consumption.

I called the objective function inside the PSO code

CostFunction = @ energy_demand; % Objective Function/Fitness function

and defined the lower and upper bounds for x1,x2,x3,x4,x5,x6 as:

VarMin = [0.8,0.8,1,0.3,0.1,0.1]; % Lower Bound of Decision Variables

VarMax = [1,1,10,0.7,0.5,0.5]; % Upper Bound of Decision Variables

I assumed MATLAB would pick the upper and lower bounds sequentially, but I keep getting the error:

Not enough input arguments.

Error in energy_demand (line 114)

Qlong =((Ta + 273.15).^4 - (To + 273.15).^4) * x1 * x2 * As * sigma;

Error in PSO (line 58)
        particle(i).Cost = CostFunction(particle(i).Position);

Error in pso1 (line 25)
out = PSO(problem, params);

The function am trying to minimize as shown below:

function RMSE = energy_demand(x1,x2,x3,x4,x5,x6)

%% greenhouse environmental parameters

environmental_input_parameters = xlsread('day and night-time environmental inputs.xlsx','T2:Z10141');

Qmeasured = xlsread('Qmeasured.xlsx');

Qreal = Qmeasured(:,2);

Ta = environmental_input_parameters(:,1); % interior temperature

slope = environmental_input_parameters(:,2); % dTa/dt

To = environmental_input_parameters(:,3); % exterior temperature

Rgo = environmental_input_parameters(:,4); % global radiation

t = environmental_input_parameters(:,5); % time series


%% Physical Constants

rho = 1.225;               %Air Density in kg.m^-3

Cp = 1006.43;              %specific heat capacity of air in J·kg^-1·K^-1

%% Greenhouse Physical Dimensions

Vol = 16315.517 ;            % Greenhouse Volume in V/m^-3

As = 5079.9263;            % surface area of the greenhouse in m^2

Ag = 3393.324;             % greenhouse ground surface area in m^2

%% Thermal Model

%Longwave Radiation

Qlong =((Ta + 273.15).^4 - (To + 273.15).^4) * x1 * x2 * As * sigma;

%% Heat exchanged between cover and the exterior of the greenhouse

Qcover = (Ta - To) * As * x3 * x4;

%% Heat Flux of Air

Qair = slope * Vol * rho * Cp;

%% Net Solar Radiation

Qsolar = Rgo * As * x5 * x6;

%% Computed greenhouse energy consumption

Qcomputed = Qair + Qlong + Qcover - Qsolar;

alpha = sum((Qcomputed(:) - Qreal(:)).^2);

N = size((Qcomputed(:)),1);

RMSE = sqrt(alpha./N);  


end

[muhammad hamza]

Same Issue im trying to minimize my objective function for boost converter which im contolling with FOPID.I need to set limits for each of 5 parameters of FOPID i need to find.

Answer 1

[Alan Weiss]

You need to write your function of one variable, say x = [x1,x2,x3,x4,x5]. For example,

function RMSE = energy_demand(x) % Here x has 5 components

For details, see Writing Scalar Objective Functions.

Alan Weiss

MATLAB mathematical toolbox documentation