A FIS(Fuzzy Inference System) object. More...

#include <fis.hpp>

Inheritance diagram for qlibs::fis::instance:

Public Member Functions
	instance ()=default

virtual	~instance ()

bool	setup (const type t, input const inputs, const size_t ni, output const outputs, const size_t no, mf const mf_inputs, const size_t nmi, mf const mf_outputs, const size_t nmo, const rules const r, const size_t n, real_t rWeights=nullptr) noexcept
	Setup and initialize the FIS instance.

template<size_t numberInputs, size_t numberOutputs, size_t numberMFinputs, size_t numberMFOutputs, size_t numberRules>
bool	setup (const type t, input(&inputs)[numberInputs], output(&outputs)[numberOutputs], mf(&mf_inputs)[numberMFinputs], mf(&mf_outputs)[numberMFOutputs], const rules *const r, real_t(&rWeights)[numberRules]) noexcept
	Setup and initialize the FIS instance.

bool	setupInput (const tag t, const real_t Min, const real_t Max) noexcept
	Setup the input with the specified tag and set limits for it.

bool	setupOutput (const tag t, const real_t Min, const real_t Max) noexcept
	Setup the output with the specified tag and set limits for it.

bool	setupInputMF (const tag io, const tag mf, const shapeMF s, const real_t *cp, const real_t h=1.0_re) noexcept
	Setup the input tag and points for the specified membership function.

bool	setupInputMF (const tag io, const tag mf, mfFunction customMfs, const real_t *cp, const real_t h=1.0_re) noexcept
	Setup the input tag and points for the specified membership function.

bool	setupOutputMF (const tag io, const tag mf, const shapeMF s, const real_t *cp, const real_t h=1.0_re) noexcept
	Setup the output tag and points for the specified membership function.

bool	setupOutputMF (const tag io, const tag mf, mfFunction customMfs, const real_t *cp, const real_t h=1.0_re) noexcept
	Set the output tag and points for the specified membership function.

bool	setInput (const tag t, const real_t value) noexcept
	Set a crisp value of the input with the specified tag.

bool	getOutput (const tag t, real_t &value) const noexcept
	Get the de-fuzzified crisp value from the output with the specified tag.

bool	setParameter (const parameter p, const paramValue x) noexcept
	Set parameters of the FIS instance.

bool	setDeFuzzMethod (deFuzzMethod m) noexcept
	Change the default de-Fuzzification method of the FIS instance.

bool	fuzzify (void) noexcept
	Perform the fuzzification operation over the crisp inputs on the requested FIS object.

bool	deFuzzify (void) noexcept
	Perform the de-Fuzzification operation to compute the crisp outputs.

bool	inference (void) noexcept
	Perform the inference process on the FIS object.

bool	setRuleWeights (real_t *rWeights) noexcept
	Set weights to the rules of the inference system.

bool	isInitialized (void) const
	Check if the FIS instance has been initialized.

	operator bool () const noexcept
	Check if the FIS instance has been initialized.

size_t	getNumberOfPoints (void) const noexcept
	Get the number of points used on Mamdani to perform the de-fuzzification proccess.

real_t	operator[] (tag outTag) const
	Get the de-fuzzified crisp value from the output with the specified tag.

instance &	operator<< (const tag &t)
	Select the input to set using with the specified tag.

instance &	operator<< (const int &value)
	The value to set the previously selected input.

instance &	operator<< (const real_t &value)
	The value to set the previously selected input.

Detailed Description

A FIS(Fuzzy Inference System) object.

The instance should be initialized using the instance::setup() method.

Constructor & Destructor Documentation

◆ instance()

qlibs::fis::instance::instance ( )

default

◆ ~instance()

virtual qlibs::fis::instance::~instance ( )

inlinevirtual

Member Function Documentation

◆ deFuzzify()

bool fis::instance::deFuzzify ( void )

noexcept

Perform the de-Fuzzification operation to compute the crisp outputs.

Precondition: The instance should have already invoked the inference process successfully with fis::inference()

Note: By default, this method, uses the Centroid method on Mamdani type FIS and weight-average on Sugeno type FIS. To change the default settings use the instance::setDeFuzzMethod() function.

Returns: true on success, otherwise return false.

◆ fuzzify()

bool fis::instance::fuzzify ( void )

noexcept

Perform the fuzzification operation over the crisp inputs on the requested FIS object.

Precondition: I/Os and fuzzy sets must be previously initialized by instance::setupInput(), instance::setupOutput(), instance::setupInputMF(), instance::setupOutputMF and instance::setup() respectively.

Returns: true on success, otherwise return false.

◆ getNumberOfPoints()

size_t qlibs::fis::instance::getNumberOfPoints ( void ) const

inlinenoexcept

Get the number of points used on Mamdani to perform the de-fuzzification proccess.

Returns: The number for points for de-fuzzification.

◆ getOutput()

bool fis::instance::getOutput	(	const tag	t,
		real_t &	value ) const

noexcept

Get the de-fuzzified crisp value from the output with the specified tag.

Parameters

[in]	t	The output tag
[out]	value	The variable where the output will be stored

Returns: The requested de-fuzzified crisp value.

◆ inference()

bool fis::instance::inference ( void )

noexcept

Perform the inference process on the FIS object.

Precondition: The instance should have already invoked the fuzzification operation successfully with instance::fuzzify()

Returns: true on success, otherwise return false.

◆ isInitialized()

bool qlibs::fis::instance::isInitialized ( void ) const

inline

Check if the FIS instance has been initialized.

Returns: true if instance has been initialized

◆ operator bool()

qlibs::fis::instance::operator bool ( ) const

inlineexplicitnoexcept

Check if the FIS instance has been initialized.

Returns: true if instance has been initialized

◆ operator<<() [1/3]

instance & qlibs::fis::instance::operator<< ( const int & value )

inline

The value to set the previously selected input.

Parameters

[in] value The crisp value to set

Returns: The fis object you invoked << upon.

◆ operator<<() [2/3]

instance & qlibs::fis::instance::operator<< ( const real_t & value )

inline

The value to set the previously selected input.

Parameters

[in] value The crisp value to set

Returns: The fis object you invoked << upon.

◆ operator<<() [3/3]

instance & qlibs::fis::instance::operator<< ( const tag & t )

inline

Select the input to set using with the specified tag.

Parameters

[in] t The input tag

Returns: The fis object you invoked << upon.

◆ operator[]()

real_t qlibs::fis::instance::operator[] ( tag outTag ) const

inline

Get the de-fuzzified crisp value from the output with the specified tag.

Parameters

[in] outTag The output tag

Returns: The requested de-fuzzified crisp value.

◆ setDeFuzzMethod()

bool fis::instance::setDeFuzzMethod ( fis::deFuzzMethod m )

noexcept

Change the default de-Fuzzification method of the FIS instance.

Parameters

[in] m The de-fuzzification method: use one of the following : centroid, bisector, mom, lom, som, wtaver, wtsum

Note: centroid, bisector, mom, lom and som only apply for a Mamdani FIS; wtaver and wtsum only apply for a Sugeno FIS.; wtaver only apply for a Tsukamoto FIS

Returns: true on success, otherwise return false

◆ setInput()

bool fis::instance::setInput	(	const tag	t,
		const real_t	value )

noexcept

Set a crisp value of the input with the specified tag.

Parameters

[in]	t	The input tag
[in]	value	The crisp value to set

Returns: true on success, otherwise return false.

◆ setParameter()

bool fis::instance::setParameter	(	const parameter	p,
		const paramValue	x )

noexcept

Set parameters of the FIS instance.

Parameters

[in]	p	The requested parameter to change/set.
[in]	x	The value of the parameter to set.

Returns: true on success, otherwise return false.

◆ setRuleWeights()

bool fis::instance::setRuleWeights ( real_t * rWeights )

noexcept

Set weights to the rules of the inference system.

Precondition: I/Os and fuzzy sets must be previously initialized by instance::setupInput(), instance::setupOutput(), instance::setInputMF(), instance::setOutputMF and instance::setup() respectively.

Parameters

[in] rWeights An array with the values of every rule weight;

Returns: true on success, otherwise return false.

◆ setup() [1/2]

bool fis::instance::setup	(	const type	t,
		fis::input *const	inputs,
		const size_t	ni,
		fis::output *const	outputs,
		const size_t	no,
		fis::mf *const	mf_inputs,
		const size_t	nmi,
		fis::mf *const	mf_outputs,
		const size_t	nmo,
		const rules *const	r,
		const size_t	n,
		real_t *	rWeights = nullptr )

noexcept

Setup and initialize the FIS instance.

Note: Default configuration : AND = Min, OR = Max, Implication = Min Aggregation = Max, EvalPoints = 100

Parameters

[in]	t	Type of inference Mamdani, Sugeno or Tsukamoto.
[in]	inputs	An array with all the system inputs as fis::input objects.
[in]	ni	The number of elements in the inputs array.
[in]	outputs	An array with all the system outputs as fis::output objects.
[in]	no	The number of elements in the outputs array.
[in]	mf_inputs	An array with all the membership functions related to the inputs. This should be an array of fis::mf objects.
[in]	nmi	The number of elements in the mf_inputs array.
[in]	mf_outputs	An array with all the membership functions related to the outputs. This should be an array of fis::mf objects.
[in]	nmo	The number of elements in the mf_outputs array.
[in]	r	The rules set.
[in]	n	Number of rules
[in]	rWeights	An array of size n were the rule strengths will be stored.

Returns: true on success, otherwise return false.

◆ setup() [2/2]

template<size_t numberInputs, size_t numberOutputs, size_t numberMFinputs, size_t numberMFOutputs, size_t numberRules>

bool qlibs::fis::instance::setup	(	const type	t,
		input(&)	inputs[numberInputs],
		output(&)	outputs[numberOutputs],
		mf(&)	mf_inputs[numberMFinputs],
		mf(&)	mf_outputs[numberMFOutputs],
		const rules *const	r,
		real_t(&)	rWeights[numberRules] )

inlinenoexcept

Setup and initialize the FIS instance.

Note: Default configuration : AND = Min, OR = Max, Implication = Min Aggregation = Max, EvalPoints = 100

Parameters

[in]	t	Type of inference Mamdani, Sugeno or Tsukamoto.
[in]	inputs	An array with all the system inputs as fis::input objects.
[in]	outputs	An array with all the system outputs as fis::output objects.
[in]	mf_inputs	An array with all the membership functions related to the inputs. This should be an array of fis::mf objects.
[in]	mf_outputs	An array with all the membership functions related to the outputs. This should be an array of fis::mf objects.
[in]	r	The rules set.
[in]	rWeights	An array of size n were the rule strengths will be stored.

Returns: true on success, otherwise return false.

◆ setupInput()

bool fis::instance::setupInput	(	const tag	t,
		const real_t	Min,
		const real_t	Max )

noexcept

Setup the input with the specified tag and set limits for it.

Parameters

[in]	t	The input tag
[in]	Min	Minimum allowed value for this input
[in]	Max	Max allowed value for this input

Returns: true on success, otherwise return false.

◆ setupInputMF() [1/2]

bool qlibs::fis::instance::setupInputMF	(	const tag	io,
		const tag	mf,
		const shapeMF	s,
		const real_t *	cp,
		const real_t	h = 1.0_re )

inlinenoexcept

Setup the input tag and points for the specified membership function.

Parameters

[in]	io	The input tag related with this membership function
[in]	mf	The user-defined tag for this membership function
[in]	s	The wanted shape/form for this membership function, can be one of the following: trimf, trapmf, gbellmf, gaussmf, gauss2mf, sigmf, dsigmf, psigmf, pimf, smf, zmf, singletonmf, concavemf, spikemf, linsmf, linzmf, rectmf, cosmf.

Note: For Sugeno FIS, an output membership function should be one of the following: constantmf, linearmf.; For Tsukamoto FIS, an output membership function should be one the following monotonic functions : tlinsmf, tlinzmf, tsmf, tzmf, tconcavemf

Parameters

[in]	cp	Points or coefficients of the membership function.
[in]	h	Height of the membership function.

Note: Height parameter h does not apply for output membership functions on Sugeno and Tsukamoto inference systems. [ 0 <= h <= 1]

Returns: true on success, otherwise return false.

◆ setupInputMF() [2/2]

bool qlibs::fis::instance::setupInputMF	(	const tag	io,
		const tag	mf,
		mfFunction	customMfs,
		const real_t *	cp,
		const real_t	h = 1.0_re )

inlinenoexcept

Setup the input tag and points for the specified membership function.

Parameters

[in]	io	The input tag related with this membership function
[in]	mf	The user-defined tag for this membership function
[in]	customMfs	Custom user-defined membership function.
[in]	cp	Points or coefficients of the membership function.
[in]	h	Height of the membership function.

Note: Height parameter h does not apply for output membership functions on Sugeno and Tsukamoto inference systems. [ 0 <= h <= 1]

Returns: true on success, otherwise return false.

◆ setupOutput()

bool fis::instance::setupOutput	(	const tag	t,
		const real_t	Min,
		const real_t	Max )

noexcept

Setup the output with the specified tag and set limits for it.

Parameters

[in]	t	The output tag
[in]	Min	Minimum allowed value for this output
[in]	Max	Max allowed value for this output

Returns: true on success, otherwise return false.

◆ setupOutputMF() [1/2]

bool qlibs::fis::instance::setupOutputMF	(	const tag	io,
		const tag	mf,
		const shapeMF	s,
		const real_t *	cp,
		const real_t	h = 1.0_re )

inlinenoexcept

Setup the output tag and points for the specified membership function.

Parameters

[in]	io	The output tag related with this membership function
[in]	mf	The user-defined tag for this membership function
[in]	s	The wanted shape/form for this membership function, can be one of the following: trimf, trapmf, gbellmf, gaussmf, gauss2mf, sigmf, dsigmf, psigmf, pimf, smf, zmf, singletonmf, concavemf, spikemf, linsmf, linzmf, rectmf, cosmf.

Note: For Sugeno FIS, an output membership function should be one of the following: constantmf, linearmf.; For Tsukamoto FIS, an output membership function should be one the following monotonic functions : tlinsmf, tlinzmf, tsmf, tzmf, tconcavemf

Parameters

[in]	cp	Points or coefficients of the membership function.
[in]	h	Height of the membership function.

Note: Height parameter h does not apply for output membership functions on Sugeno and Tsukamoto inference systems. [ 0 <= h <= 1]

Returns: true on success, otherwise return false.

◆ setupOutputMF() [2/2]

bool qlibs::fis::instance::setupOutputMF	(	const tag	io,
		const tag	mf,
		mfFunction	customMfs,
		const real_t *	cp,
		const real_t	h = 1.0_re )

inlinenoexcept

Set the output tag and points for the specified membership function.

Parameters

[in]	io	The output tag related with this membership function
[in]	mf	The user-defined tag for this membership function
[in]	customMfs	Custom user-defined membership function.
[in]	cp	Points or coefficients of the membership function.
[in]	h	Height of the membership function.

Note: Height parameter h does not apply for output membership functions on Sugeno and Tsukamoto inference systems. [ 0 <= h <= 1]

Returns: true on success, otherwise return false.

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ instance()

◆ ~instance()

Member Function Documentation

◆ deFuzzify()

◆ fuzzify()

◆ getNumberOfPoints()

◆ getOutput()

◆ inference()

◆ isInitialized()

◆ operator bool()

◆ operator<<() [1/3]

◆ operator<<() [2/3]

◆ operator<<() [3/3]

◆ operator[]()

◆ setDeFuzzMethod()

◆ setInput()

◆ setParameter()

◆ setRuleWeights()

◆ setup() [1/2]

◆ setup() [2/2]

◆ setupInput()

◆ setupInputMF() [1/2]

◆ setupInputMF() [2/2]

◆ setupOutput()

◆ setupOutputMF() [1/2]

◆ setupOutputMF() [2/2]