Model proposition for predicting sustainability classes using multicriteria decision support and artificial intelligence

Couto, Ayrton Benedito Gaia do; Rangel, Luís Alberto Duncan

doi:10.1590/1806-9649-2022v29e6922

Sumário Anterior Atual Seguinte

Sumário

ORIGINAL ARTICLE • Gest. Prod. 29 • 2022 • https://doi.org/10.1590/1806-9649-2022v29e6922 linkcopy

Model proposition for predicting sustainability classes using multicriteria decision support and artificial intelligence

Proposição de modelo para a predição de classes de sustentabilidade por meio de apoio à decisão multicritério e inteligência artificial

Authorship SCIMAGO INSTITUTIONS RANKINGS

The current study proposes a novel prediction model of sustainability classes for electricity distribution companies in Brazil, based on sustainability indicators, aiming at a more effective risk management for a certain company among their competitors. Because such indicators are based on quantitative and qualitative measures and are very likely to incur imprecisions in their measures, the model to be proposed is based on a Multicriteria Decision Support, Rough Sets Theory, which allows the mathematical treatment of those imprecisions, and Artificial Intelligence, in this case, Machine Learning by rules inference. Consequently, decision tables are generated with condition attributes, sustainability indicators, and decision attributes, sustainability classes: high, medium or low. As a result, it is possible to predict sustainability classes based in temporal series of indicators and rules inference from decision tables, using RoughSets package in R and the jMAF software, demonstrating the use of five rule generation algorithms and their respective accuracies.

Keywords:
Forecasting; Multicriteria decision; Rough sets theory; Artificial intelligence; Sustainability; Risk management

PHASE 1	PHASE 2	PHASE 3	PHASE 4	PHASE 5	PHASE 6	PHASE 7	PHASE 8
Set keywords for the search	Define bases aligned with research	Perform title, abstract and keyword research of publications in each database	Count the number of publications per axis and base in order to build the initial bibliographic portfolio (BP0)	Raise number of BP1 citations on Google Scholar	Select most cited publications (up to 85% of total citations)	Select least cited publications (remaining 15% of total citations)	Gather publications BP2, BP3 and BP4
Define research axes; assemble boolean expressions (“and”; “or”)	Set search filters (time period, subject, publication type, etc.)	Review whether searched keywords adhere to the keywords of the publications	Select publications adhering to the research (reading the title); exclude repeated posts (BP1)	Sort publications in descending order of citations	Select publications adhering to the research (abstract reading) (BP2)	Select publications in the last 2 years adhering to the research (abstract reading) (BP3);	Select publications adhering to the research (full reading) to assemble the final bibliographic portfolio (BP5)
				Sort publications in descending order of citations		Select remaining publications that have authors in BP2 and adherent to the research (abstract reading) (BP4)

Author(s)	Objective	Method(s)	Result(s)
Ahmad et al. (2021)	A baseline that allows researchers and readers to compare their AI efforts, new state-of-the-art applications and global roles in policymaking of energy industry	AI techniques	The energy industry, utilities, power system operators, and independent power producers may need to focus more on AI technologies if they want meaningful results to remain competitive.
Ahmadi et al. (2022)	Predicting energy consumption patterns in urban buildings	Deep neural network with fuzzy wavelets	This study shows that the presented method provides high-performance prediction at a lower level of complexity.
Al-Barakati et al. (2022)	Evaluating the renewable energy sources	Extended interval-valued Pythagorean fuzzy WASPAS	The evaluation results showed that the wind energy with a maximum assessment score degree using the proposed method was found the best option for selecting renewable energy sources over diverse criteria.
Buțurache & Stancu (2022)	Building energy consumption prediction	Neural-based models	Neural-based models possess the capability of learning and generalizing from different datasets having different patterns.
Caiado et al. (2017)	Proposing a novel model for solving decision-making problems in the evaluation of electrical energy companies	TOPSIS method	Robust evaluation and ranking of the energy companies with respect to the observed aspects of sustainability.
Chamandoust et al. (2020)	Performance assessment of smart hybrid energy system (SHES)	Shuffled frog leaping algorithm (SFLA)	Optimal scheduling of SHES with acceptable levels of operation costs, emission pollution and customer satisfaction.
Colla et al. (2020)	Critically review of multi-disciplinary KPIs, allowing a holistic comparison across different types of energy projects	A structured evaluation framework based on the identified set of indicators	Integrated framework and a fairer assessment of competing energy projects by relevant stakeholders.
Daugavietis et al. (2022)	A Comparison of Multi-Criteria Decision Analysis Methods for sustainability assessment of District Heating systems	WSM, TOPSIS, PROMETHEE, ELECTRE and DEA methods	The results of sensitivity analysis along with literature investigation shows that all methods are suitable for sustainability analyses of District Heating systems while also having differences in the calculation process and in the interpretation of results.
Ervural et al. (2018a)	Energy investment planning	TOPSIS and AHP	The renewable energy investment plan of a power company in Turkey is evaluated with a newly developed integral approach.
Ervural et al. (2018b)	Energy planning	ANP, fuzzy TOPSIS and SWOT analysis	Integrated framework for the Turkey’s energy sector to prioritize alternative energy strategies.
Kwakkel & Pruyt (2013)	An approach for model-based foresight under deep uncertainty	Exploratory Modeling and Analysis	Multiplicity assessment of deep uncertainties in the analysis of decision-making problems in the electricity sector.
Panchal et al. (2022)	A novel structured framework for analyzing sustainable operational performance‐related issues of ash handling unit (AHU) under vague/uncertain information	Integrated fuzzy lambda-tau and fuzzy multicriteria decision‐making methods	The current work discussed a framework for evaluating performance issues of the thermal power industry for sustainable and environmental friendly operation for emission control.
Qi et al. (2020)	Defining a set of criteria and dimensions for analyzing the corporate governance-based strategic	IVIF DEMATEL; IVIF VIKOR	Extending investigations on corporate governance and sustainable production in energy industry.
Rigo et al. (2020)	Systematic literature review of renewable energy problems associated with MCDM methods	MCDM methods	Improving their ability to choose the proper MCDM methods to solve energy problems.
Rolnick et al. (2022)	How machine learning can be a powerful tool in reducing greenhouse gas emissions and helping society adapt to a changing climate	Machine Learning techniques	Climate change solution domains with selected areas of ML that are relevant.
Sahabuddin & Khan (2021)	Energy sector’s sustainability assessment	Multi-criteria decision analysis methods	The analysis revealed that COPRAS is the most robust MCDA method, followed by WPM.
Tajbakhsh & Shamsi (2019)	Analyzing a comprehensive objective function in the facility location context in the presence of variable costs, fixed costs, and sustainability consideration	Double bootstrap data envelopment analysis	The proposed approach substantially diminishes greenhouse gas emissions at the cost of slight increases in total expense.
Vargas-Solar et al. (2022)	Providing data analytical tools for metering household energy consumption and CO₂ footprint under different perspectives	GREENHOME environment toolkit	The article reports on experiments conducted for modelling and forecasting energy consumption and CO₂ footprint in the context of the Triple-A European project.
Wang et al. (2021)	Using sustainable performance prediction in data-scarce scenarios	Integrated energy efficiency system	Strengthening the real-time control of integrated energy projects and for effectively promoting the sustainable development of the integrated energy industry.
Wanke et al. (2020)	Improving the technical efficiency of the energy industry in China	Bayesian stochastic frontier analysis	Increases in the efficiency of the Chinese energy industry can be achieved by increasing the level of inventories and fixed assets.
Zhou et al. (2019)	Evaluation of sustainable energy investment projects	Balanced Scorecard, IT2 Fuzzy DEMATEL, IT2 Fuzzy QUALIFLEX	Which issues are effective in financial institutions' lending process of large-scale energy projects.

Store	E	Q	L	P
1	High	Good	No	Profit
2	Medium	Good	No	Loss
3	Medium	Good	No	Profit
4	None	Medium	No	Loss
5	Medium	Medium	Yes	Loss
6	High	Medium	Yes	Profit

Indicator Description	Reference	Dimension	Unit	Initials	(G)ain/(C)ost
Direct energy consumption per primary energy source	ANEEL (2015); GRI (2000): EN3	Environmental	kWh	IA1	C
Total water consumption	ANEEL (2015); GRI (2000): EN8	Environmental	m³	IA2	C
Annual volume of greenhouse gases (CO2, CH4, N2O, HFC, PFC, SF6) emitted into the atmosphere (in tons of equivalent CO2)	ANEEL (2015); GRI (2000): EN16, EN17, EN18	Environmental	t CO2	IA3	C
Total disposal of water, by quality and destination	ANEEL (2015); GRI (2000): EN21	Environmental	m³	IA4	C
Quantity of residues contaminated by PCB (Ascarel) destined	ANEEL (2015); GRI (2000): EN24	Environmental	t	IA5	C
Number of end users served	ANEEL (2015); GRI (2000):EU3	Economic	un	IE1	G
Number of end users served - Free	ANEEL (2015); GRI (2000): EU3	Economic	un	IE2	G
Energy bought (GWh)	ANEEL (2015); GRI (2000): EU10	Economic	GWh	IE3	C
Energy sold (GWh)	ANEEL (2015); GRI (2000): EU3, 2.7	Economic	GWh	IE4	G
Substations (in units)	ANEEL (2015); GRI (2000): EU1	Economic	un	IE5	G
Number of own employees	ANEEL (2015); GRI (2000): LA1	Social	un	IS5	G
Equivalent Duration Interruption per Consumer Unity “DEC”, company-wide - Calculated value	ANEEL (2015); GRI (2000): EU29	Social	h	IS1	C
Equivalent Duration Interruption per Consumer Unity “DEC”, company-wide - Limit	ANEEL (2015); GRI (2000): EU29	Social	h	IS2	C
Equivalent Frequency Interruption per Consumer Unity “FEC”, company-wide - Calculated value	ANEEL (2015); GRI (2000): EU28	Social	un	IS3	C
Equivalent Frequency Interruption per Consumer Unity “FEC”, company-wide - Limit	ANEEL (2015); GRI (2000): EU28	Social	un	IS4	C

Indicator Description	Reference	Unidade	Initials	(G)ain/(C)ost
Preferencial acts broadcaster.	Pereira & Cândido (2020)	0 or 1	IG01	G
Preferencial shareholders have the right to vote in relevant matters.	Pereira & Cândido (2020)	0 or 1	IG02	G
Disclosure mechanisms about the deliberated themes in the assemblies.	Pereira & Cândido (2020)	0 or 1	IG03	G
Prohibition of loans and guarantees in favour of the controller, the managers and other related parties.	Pereira & Cândido (2020)	0 or 1	IG04	G
Existence of a channel dedicated to anonymous communications that is intended to receive reports, complaints and suggestions.	Pereira & Cândido (2020)	0 or 1	IG05	G
Annual and/or sustainability report with accessibility for people with disabilities.	Pereira & Cândido (2020)	0 or 1	IG06	G
Programs for education on sustainability and the audiences reached.	Pereira & Cândido (2020)	0 or 1	IG07	G
Adherence of voluntary commitments related to sustainable development.	Pereira & Cândido (2020)	0 or 1	IG08	G
Existence of a sustainability committee.	Pereira & Cândido (2020)	0 or 1	IG09	G
Independent audit opinion.	Pereira & Cândido (2020)	0 or 1	IG10	G
Commitment with the fight against corruption that encompasses the internal audience.	Pereira & Cândido (2020)	0 or 1	IG11	G
Commitment with the fight against corruption of company partners.	Pereira & Cândido (2020)	0 or 1	IG12	G
Organizations’ governance structure, including Committees under the highest governance body responsible by specific tasks, such as strategic configuration or organizational supervision. Describe the mandate and the composition (including number of independent and/or non-executive members) of such committees and assign any direct responsibility by economic, social and environmental performance matters.	GRI (2000)	0 or 1	IG13	G
For organizations that have a unitary council structure, indicate the number of members from the highest governance body who are independent and/or non-executive members. Declare how the organization defines “independent” and “non-executive”. This element applies only to organizations that have unitary council structures.	GRI (2000)	0 or 1	IG14	G
Mechanisms that enable shareholders and employees to supply recommendations or orientations for the highest governance body. It includes a reference to processes related to: • The use of shareholder resolutions or other mechanisms in order to allow the shareholder minority to express opinions to the highest governance body; and • Report and consult employees about work relationships with the representative bodies, such as a “work council” level of organization, and representation of employees of the highest governance body. Identify topics related to economic, environmental, and social performance aspects generated by these mechanisms throughout the period covered by the report.	GRI (2000)	0 or 1	IG15	G
Processes in effect for the highest governance body in order to guarantee that conflicts of interest are avoided.	GRI (2000)	0 or 1	IG16	G
Process to determine the qualifications and experience from the members of the highest level of governance body in order to guide the organization regarding the economic, environmental, and social strategies.	GRI (2000)	0 or 1	IG17	G
Mission statements developed internally or values, codes of conduct and relevant principles for the economic, environmental, and social performance and the status of its implementation. Explain the level in which these: • Are applied in the entire organization in different regions and departments/units; and • Relate to the internationally accepted standards.	GRI (2000)	0 or 1	IG18	G
Procedures from the highest governance body to supervise the organization: economic, environmental, and social performance management, including relevant risks and opportunities for adherence or conformity to internationally accorded standards, code of conduct and principles. It includes the frequency with which the highest governance body evaluated the sustainability performance.	GRI (2000)	0 or 1	IG19	G
Processes to evaluate the highest governance, particularly regarding the economic, environmental and social performance.	GRI (2000)	0 or 1	IG20	G

Indicator Description	Initials	Keywords
Preferencial acts broadcaster.	IG01	preferencials
Preferencial shareholders have the right to vote in relevant matters.	IG02	preferencials
Disclosure mechanisms about the deliberated themes in the assemblies.	IG03	disclosure
Prohibition of loans and guarantees in favour of the controller, the managers and other related parties.	IG04	prohibition
Existence of a channel dedicated to anonymous communications that is intended to receive reports, complaints and suggestions.	IG05	channel
Annual and/or sustainability report with accessibility for people with disabilities.	IG06	report; sustainability
Programs for education on sustainability and the audiences reached.	IG07	education
Adherence of voluntary commitments related to sustainable development.	IG08	adherence
Existence of a sustainability committee.	IG09	committee
Independent audit opinion.	IG10	independent audit
Commitment with the fight against corruption that encompasses the internal audience.	IG11	corruption
Commitment with the fight against corruption of company partners.	IG12	corruption
Organizations’ governance structure, including Committees under the highest governance body responsible by specific tasks, such as strategic configuration or organizational supervision. Describe the mandate and the composition (including number of independent and/or non-executive members) of such committees and assign any direct responsibility by economic, social and environmental performance matters.	IG13	committee; independent; independence
For organizations that have a unitary council structure, indicate the number of members from the highest governance body who are independent and/or non-executive members. Declare how the organization defines “independent” and “non-executive”. This element applies only to organizations that have unitary council structures.	IG14	independent; independence
Mechanisms that enable shareholders and employees to supply recommendations or orientations for the highest governance body. It includes a reference to processes related to: • The use of shareholder resolutions or other mechanisms in order to allow the shareholder minority to express opinions to the highest governance body; and • Report and consult employees about work relationships with the representative bodies, such as a “work council” level of organization, and representation of employees of the highest governance body. Identify topics related to economic, environmental, and social performance aspects generated by these mechanisms throughout the period covered by the report.	IG15	recommendations; recommendation; orientation; orientations
Processes in effect for the highest governance body in order to guarantee that conflicts of interest are avoided.	IG16	conflict; interest
Process to determine the qualifications and experience from the members of the highest level of governance body in order to guide the organization regarding the economic, environmental, and social strategies.	IG17	qualifications; experience
Mission statements developed internally or values, codes of conduct and relevant principles for the economic, environmental, and social performance and the status of its implementation. Explain the level in which these: • Are applied in the entire organization in different regions and departments/units; and • Relate to the internationally accepted standards.	IG18	declarations; declaration; values; conduct; principles
Procedures from the highest governance body to supervise the organization: economic, environmental, and social performance management, including relevant risks and opportunities for adherence or conformity to internationally accorded standards, code of conduct and principles. It includes the frequency with which the highest governance body evaluated the sustainability performance.	IG19	supervision; supervise; risks; adherence; conformity; standars; standard; performance; sustainability
Processes to evaluate the highest governance, particularly regarding the economic, environmental and social performance.	IG20	evaluation; evaluate

Year	Company	InA	InE	InS	InG	Class A	Class E	Class S	Class G	In Su	Class Su
2019	E01	0.3579	0.8804	0.9320	0.3333	M	H	H	M	0.6259	M
	E02	0.7087	0.6271	0.7807	0.6667	H	M	H	H	0.6958	H
	E03	0.6135	0.6458	0.7865	1.0000	M	M	H	H	0.7615	H
	E04	0.5101	0.7868	0.9566	0.8333	M	H	H	H	0.7717	H
	E05	0.4298	0.6691	0.8427	0.5000	M	H	H	M	0.6104	M
	E06	0.2934	0.7630	0.9051	0.5000	L	H	H	M	0.6154	M
	E07	0.3077	0.7778	0.3908	0.5000	L	H	M	M	0.4941	M
	E08	0.3578	0.5806	0.8075	1.0000	M	M	H	H	0.6865	H
	E09	0.4761	0.4555	0.8119	0.5000	M	M	H	M	0.5609	M
	E10	0.4493	0.6715	0.8557	1.0000	M	H	H	H	0.7441	H
2018	E01	0.4255	0.8646	0.9663	0.3333	M	H	H	M	0.6474	M
	E02	0.7267	0.5913	0.7833	0.6667	H	M	H	H	0.6920	H
	E03	0.6458	0.6352	0.7518	1.0000	M	M	H	H	0.7582	H
	E04	0.5448	0.7604	0.9389	0.8333	M	H	H	H	0.7693	H
	E05	0.4502	0.6211	0.7931	1.0000	M	M	H	H	0.7161	H
	E06	0.4275	0.3845	0.8698	0.5000	M	M	H	M	0.5455	M
	E07	0.4878	0.4053	0.3780	0.1667	M	M	M	L	0.3594	M
	E08	0.4411	0.6381	0.8306	1.0000	M	M	H	H	0.7274	H
	E09	0.5364	0.5886	0.8473	0.5000	M	M	H	M	0.6181	M
	E10	0.4757	0.6904	0.8856	1.0000	M	H	H	H	0.7629	H
2017	E01	0.6396	0.8502	0.9461	0.3333	M	H	H	M	0.6923	H
	E02	0.6730	0.5228	0.7449	0.6667	H	M	H	H	0.6518	M
	E03	1.0000	0.5682	0.6544	1.0000	H	M	M	H	0.8056	H
	E04	0.5287	0.7141	0.9360	0.8333	M	H	H	H	0.7530	H
	E05	0.5594	0.5383	0.3198	0.0000	M	M	L	L	0.3544	M
	E06	0.6396	0.5720	0.7372	0.6667	M	M	H	H	0.6538	M
	E07	0.5999	0.6250	0.3685	0.5000	M	M	M	M	0.5234	M
	E08	0.6273	0.6936	0.6907	1.0000	M	H	H	H	0.7529	H
	E09	0.5879	0.5877	0.8214	0.5000	M	M	H	M	0.6243	M
	E10	0.6936	0.5887	0.7809	1.0000	H	M	H	H	0.7658	H

Year	Company	IA1	IA2	IA3	IA4	IA5	IE1	IE2	IE3	IE4	IE5	IS1	IS2	IS3	IS4	IS5	IG	Class Su
2019	E01	37827660	157795	2548765	126236	80	8537040	1441	47463	25562	409	10.62	10.53	5.05	7.24	4203	0.3333	M
	E02	22979922	81631	141311	235301	17	3049218	1149	19087	14454	170	10.86	11.21	7.52	8.94	3404	0.6667	H
	E03	158055	95513	465755	191577	68	6103473	359	22132	17166	299	12.19	16.46	5.91	8.31	4212	1.0000	H
	E04	22721951	100022	214073	212786	36	4713240	1389	26163	19784	374	9.11	10.06	6.02	7.67	4964	0.8333	H
	E05	24315161	98134	2349742	138027	51	8273647	409	39019	15579	407	10.16	12.43	6.73	7.63	4012	0.5000	M
	E06	19378319	132500	1470540	245000	73	8281213	1426	31595	15425	265	10.69	10.32	5.14	8.06	4055	0.5000	M
	E07	10949199	124483	2116566	218290	76	8364068	1031	28848	23211	255	10.02	28.33	24.32	28.22	1120	0.5000	M
	E08	25211945	107962	1932805	186122	62	3741228	515	40940	24405	281	10.82	14.57	7.24	8.38	4679	1.0000	H
	E09	18723750	90508	958443	145935	37	255045	454	34106	16216	302	11.59	14.82	7.16	7.33	4476	0.5000	M
	E10	16948845	90964	489174	168420	56	7553881	569	39476	17516	372	11.36	12.66	6.19	7.46	4446	1.0000	H
2018	E01	41244038	130514	3391332	104411	80	8409044	1138	45245	25271	400	10.05	10.58	5.06	7.26	4530	0.3333	M
	E02	4272084	59104	159462	171306	50	2975190	947	18572	14251	168	10.68	11.72	7.33	9.39	3333	0.6667	H
	E03	101955	84277	1843294	136922	74	5993104	302	21128	16522	341	14.44	14.5	6.44	8.42	3729	1.0000	H
	E04	10474221	99614	244660	132672	73	4637804	1121	25751	19594	369	10.31	10.44	6.22	8.01	5364	0.8333	H
	E05	39339116	84471	1316268	133178	78	5434772	352	28064	17782	370	11.47	13.56	7.32	7.60	3610	1.0000	H
	E06	39847252	121715	2927338	117504	71	312745	118	35326	22022	199	13.10	12.88	5.77	7.87	5217	0.5000	M
	E07	8726764	87185	571635	170945	58	1503806	194	14609	4452	200	11.86	28.33	24.32	28.22	1120	0.1667	M
	E08	15002688	117408	1057711	120331	74	6627398	890	27592	15994	183	11.05	13.96	6.51	7.49	4016	1.0000	H
	E09	26809396	59417	1039882	125106	72	255045	1006	42668	18277	385	12.50	13.73	7.23	7.35	5278	0.5000	M
	E10	30247811	109055	297812	155485	80	4990851	690	21375	19863	313	10.91	10.46	6.33	7.90	4235	1.0000	H
2017	E01	37476519	76060	133286	60848	90	8337594	989	33173	25091	404	10.83	10.80	5.44	7.60	4394	0.3333	H
	E02	37476519	64789	133286	60848	80	2899170	823	18290	14132	167	12.35	12.11	8.35	10.00	3298	0.6667	M
	E03	71430	26754	133286	60848	76	5900258	246	21383	16264	341	19.83	15.01	8.23	8.84	2897	1.0000	H
	E04	30361008	96237	316296	60848	81	4560493	991	25983	19743	369	10.46	10.88	6.83	8.53	5746	0.8333	H
	E05	30911709	61688	281906	60848	86	211000	203	8400	22003	236	18.74	22.47	25.94	32.04	652	0.0000	M
	E06	7007485	61085	157249	60848	84	4804650	806	24801	17659	173	14.92	12.49	7.77	9.58	3596	0.6667	M
	E07	32264140	78233	190602	60848	80	5958615	837	32437	19500	214	13.52	28.33	24.32	28.22	1120	0.5000	M
	E08	22379199	61181	179791	60848	80	6676763	804	30723	21741	289	17.00	12.90	7.95	9.67	3052	1.0000	H
	E09	9169908	63299	234148	60848	81	255045	817	30250	22332	370	11.85	12.19	8.13	9.08	4781	0.5000	M
	E10	22435685	28000	246230	60848	79	8200036	284	18484	14561	258	18.13	10.88	6.99	9.52	4357	1.0000	H

Year	Company	#	Class Su	Rule inference algorithms
Year	Company	#	Class Su	AQRules	CN2Rules	indiscernibilityBasedRules	LEM2Rules
2019	E05	5	M	M	M	H	M
2019	E08	8	H	M	H	H	H
2018	E01	11	M	M	M	M	M
	E02	12	H	H	H	H	H
	E05	15	H	H	H	H	H
	E07	17	M	H	H	H	H
	E09	19	M	M	M	H	M
2017	E03	23	H	H	H	M	M
	E08	28	H	M	H	H	M
	E10	30	H	H	H	H	M
			Accuracy	0.7	0.9	0.6	0.6

location_on

Universidade Federal de São Carlos Departamento de Engenharia de Produção , Caixa Postal 676 , 13.565-905 São Carlos SP Brazil, Tel.: +55 16 3351 8471 - São Carlos - SP - Brazil
E-mail: gp@dep.ufscar.br

rss_feed Acompanhe os números deste periódico no seu leitor de RSS