Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

Contents lists available at ScienceDirect

Journal of Open Innovation: Technology, Market,

and Complexity
journal homepage: www.sciencedirect.com/journal/journal-of-open-innovation-technology-
market-and-complexity

An approach to the integral optimization of investment portfolios

Rafael Guillermo García-Cáceres a, *, Franklin Ignacio Páez-Rivera b, Bernarda Aldana-Gómez c,
Ernesto Acosta-Gempeler c, John Wilmer Escobar-Velásquez d
a
School of Industrial Engineering, Universidad Pedagógica y Tecnológica de Colombia - UPTC, Sogamoso, Colombia
b
Industrial Engineering, Pontificia Universidad Javeriana, Bogotá, Colombia
c
Escuela Colombiana de Ingeniería Julio Garavito, Bogotá, Colombia
d
Accounting and Finance Department, Universidad del Valle, Bogotá, Colombia

A R T I C L E I N F O A B S T R A C T

Keywords: A comprehensive approach to the optimization of financial portfolios is provided by the Integral Analysis Method
Investment Portfolio (IAM), which comprises four steps: The description of the problem is followed by three mathematical steps,
Profitability of an Investment Portfolio namely, cardinal analysis, ordinal analysis and integration analysis, which allow simultaneously involving
Risk of an Investment Portfolio
ordinal and cardinal variables in an optimization problem. The cardinal analysis uses the L_1 risk model, which is
Minimum Expected Profitability Rate
Integral Analysis Method
enriched by the inclusion of a stochastic constraint on the successful performance forecast attributed to each of
Integral optimization the shares that make up the investment portfolio. This bound, however, does not alter the linear nature of the
model. The Ordinal analysis includes expert opinions on the reputation of each company as a qualitative cri­
terion. This completes two of the most relevant aspects of the investment portfolio optimization problem: success
probability (assessed through the cardinal analysis) and business reputation (assessed through the ordinal
analysis). By resorting to Stochastic Multicriteria Acceptability Analysis, the integration analysis allows
obtaining indicators that jointly evaluate the variables resulting from the two previous analyses. As a validation
process, IAM was applied to simulated data.

Nevertheless, the IAM’s cardinal analysis can be supported on any investments can be optimized when it is balanced with the systematic
other model which, at this point, ends up being enriched by the integral risk of the assets in question. However, Clyman (1995) points out that
optimization approach. It is worthwhile noting, at this point, that the L1 traditional approaches do not always give good results because the
risk-model (Konno and Yamazaki, 1991), which is detailed below. is capital percentage that should ideally be invested in a given share of a
complemented in the current work with a novel constraint associated to fixed portfolio is affected by exogenous factors. These include, for
marginal expert opinions on the relative importance of each financial example, the probability of revenue increment over time, among others,
asset. which have to be taken into consideration through expert opinions.
The current literature review revealed that investment portfolio
1. Introduction optimization has been approached through uni and multi-objective
treatments, resorting to parameters that range from deterministic to
Generally speaking, “all the decisions determining the future activ­ stochastic. In choosing the most favorable options for an optimum
ities of an individual are known as a portfolio” (Sharpe, 1999). However, portfolio combination, these research works usually look forward to
this same author states that, according to the academic perspective, the maximizing profit and/or minimizing risk within acceptable CPU times.
decisions affecting a portfolio are specifically related to the selection of Only few works have focused on the qualitative aspects of the
an appropriate set of investments. An investment portfolio might include problem, while none of them has integrated both quantitative and
shares, bonds, investment funds, exchange traded funds, cash in­ qualitative aspects. This clearly points at a tacit deficiency of the liter­
vestments, real estate property and alternative assets. Constituting a ature when it comes to optimizing the problem of investment portfolios.
portfolio basically implies the selection of a series of shares in which to Along these lines, non-probabilistic approaches involving expert
invest specific amounts of money. The profit obtained from these opinions from stock brokers and investors about risk and profitability

* Corresponding author.
E-mail address: [email protected] (R.G. García-Cáceres).

https://doi.org/10.1016/j.joitmc.2024.100235
Received 29 November 2023; Received in revised form 31 January 2024; Accepted 10 February 2024
Available online 12 February 2024
2199-8531/© 2024 The Authors. Published by Elsevier Ltd on behalf of Prof JinHyo Joseph Yun. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).
R.G. García-Cáceres et al. Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

have been certainly limited, inasmuch as they have to do with the emotional states.
intuitive notion of the result of a given investment within the planning In sum, the literature reports two different approaches to the solution
horizon. The current work proposes a more realistic modeling of the of the problem: On the one hand, the quantitative modeling of qualita­
problem by technically incorporating two decision-making factors that tive aspects is coupled to optimization solution processes. On the other
are new in the current context: expert opinions on the relative impor­ hand, MCDM techniques are applied, which are not far from the notion
tance of the assets in question (quantitative aspect) and the reputation of of optimization.
said assets (qualitative aspect). The modeling technique employed for As it can be deduced, none of these efforts has come close to the
this integral investment-portfolio optimization corresponds to a specific concept of integral optimization, which indicates the separate organi­
application of the Integral Analysis Method – IAM (García-Cáceres et al., zation of the quantitative and qualitative aspects of the problem, to be
2009). This procedure, which comes to be an alternative approach to the followed by their technical integration in an optimization context. This
problem, has been specifically designed to deal with complex optimi­ is precisely the notion that supports IAM, which is applied by the present
zation contexts involving both qualitative and quantitative attributes. work in pursuing a better approach to the problem of investment port­
The value of this contribution results from integrating human and folios. The next section details the corresponding methodological
computational potentials. In real investment portfolio decision making display.
contexts these potentials usually take separate ways: Sometimes they
correspond to the intuition and background of decision makers, while in 3. IAM application
other cases it is decision making support that takes the lead. Few works,
like the present one, have integrated both human intuition and machine IAM is featured by its capability of including quantitative and
computational force. This is illustrated through a detailed computa­ qualitative aspects in optimization contexts. It resorts to theoretical
tional example following the methodological steps of IAM to facilitate a support from both optimization techniques and those deterministic and
clear understanding of the process. ordinal versions of Stochastic Multicriteria Acceptability Analysis
(SMAA) that fit its needs. Hence, the use of these methodological tools is
2. Background actually embedded in that of IAM, which, as a consequence, comes to be
a method that resorts to other methods. IAM comprises four stages: 1.
The investment portfolio problem has been treated through not only Definition of the problem, 2. Cardinal analysis, 3. Ordinal Analysis, and
decision-making techniques such as Fuzzy Multi Attribute Group 4. Integration Analysis. A more detailed description of the technique can
(Debashree and Debajani, 2011; Tzy-Yuan et al., 2006), but also math­ be found in its original introductory publication (García-Cáceres et al.,
ematical programming procedures, as is the case of Multi–Objective 2009). These stages are illustrated here through an example based on
(Jana et al., 2009), Linear (Konno and Yamazaki, 1991) and Dynamic simulated data, which is developed in the lines that follow.
(Celikyurt and Özekici, 2007) stochastic programming. The research
targets aimed in these works include maximum revenue and best in­ 3.1. Definition of the problem
vestment selection, portfolio profit maximization, and risk tolerance
impact, among others. In addition, research on the topic has also One of the most popular and reputed portfolio optimization methods
contemplated portfolio-risk free rates; risk constraints, i.e., constants is the model of Markowitz (1952), which measures the distance between
that should be lesser or equal to the maximum bound of the investor’s variables in the L2 space, i.e., it calculates the integrals of the square
admissible risk; available capital assigned to the selected financial assets values of the differences between variables (standard deviation). Due to
(which is usually invested in its entirety); proportion of the amount to be its computational costs, this model has not been frequently used when it
invested that is assigned to each financial asset (larger than or equal to comes to optimizing large portfolios. Instead, the current work resorts to
zero); and revenue rates of the financial assets (usually associated to the the model proposed by Konno and Yamazaki (1991), which considerably
amounts to be invested), which should be larger or equal to the mini­ reduces computational costs without losing its capability to handle
mum revenue rate demanded by the investor. practical instances of the problem. This model operates in the L1 space,
For its part, IAM has been applied in diverse contexts such as wherein distances are calculated through integrals of the absolute values
container loading (García-Cáceres et al., 2011), production program­ of the differences between variables (absolute mean deviation), thus
ming (González-Neira et al., 2014) and facility location (García-Cáceres leading to a linear model which, under strict assumptions, can be ho­
et al., 2009), the latter corresponding to the initial introduction of the mologated to that of Markowitz (1952).
method. Table 1 presents a detailed literature review of the problem of The current work focuses on the integral solution of the problem. For
investment portfolios. this purpose, it includes a technical procedure to incorporate expert
The current literature on investment portfolios has mainly focused opinions and optimize the investment portfolio. This is, indeed, a central
on logarithmic efficiency to solve large scale optimization mathematical aspect of the problem, inasmuch as the opinion of specialists is
programming models. However, this research line is not explored in the frequently the most important decision-making consideration in these
current work which, instead, partially echoes a recent alternative cases, if not the only one. In this sense, the present work contributes to
investigation path. In effect, new mathematical programming models obtaining better solutions to the current problem when compared to
are being developed that integrate relevant quantitative aspects of the those usually attained in practice, as it is generally described in the
problem, related to both the objective and the constraints. Another literature.
research line which has been significantly developed is the use of MCDM
techniques to support investment portfolio decisions (cf. Kandakoglu 3.2. Cardinal analysis
et al., 2023). Qualitative considerations, in turn, have been addressed by
means of multi-objective functions which, in some cases, have included The current work resorts to the L1 risk-model (Konno and Yamazaki,
fuzzy decision environments. There is also a research line based on 1991), which is actually a variation of the Markowitz model, a classical
different types of statistical analysis, which is capable of identifying reference for the topic of investment portfolios that has been broadly
patterns and trends as well as making predictions. This approach allows applied with many variations up to date (Xie, 2021). Nevertheless, the
including both types of variables, as it is done in the work of Hung et al. IAM’s cardinal analysis can be supported on any other model which, at
(2024), which includes elements of sentiment analysis. The latter refers this point, ends up being enriched by the integral optimization
to the use of natural language processing (NLP), text mining, and com­ approach. It is worthwhile noting, that the L1 risk-model (Konno and
puter linguistics to identify and extract subjective information from the Yamazaki, 1991), which is detailed below. is complemented in the
source material, can facilitate the understanding of textually conveyed current work with a novel constraint associated to marginal expert

2
R.G. García-Cáceres et al. Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

Table 1
Literature Review.
A project Portfolio selection procedure is combined with a series of Criterion sorting or ranking methods, resulting in a new methodological proposal,
which allows prioritizing different projects based on qualitative and quantitative aspects. Applicable in different contacts, This novel methodology is
Barbati et al. (2023)
named priority-based portfolio selection open parenthesis pvps close parenthesis. In the current work the methodology is tested in the context of the
ReUse of the cultural heritage of historical cities.

A portfolio selection model is proposed in which the future risks and return rates of investment funds are represented by fuzzy triangular numbers. An
optimization method making use of fuzzy models is implemented, intending to determine the optimum investment proportion by means of a
Chen and Huang (2009)
conglomerate analysis that classifies the great many investment capital funds on the basis of four evaluation indexes: return rates, standard deviation,
rotation rate and Treynor’s index. In this way, the investors are expected to have better decision-making support.
A portfolio selection problem accommodating qualitative views is formulated. Built on the Black-Litterman model for portfolio selection, linear
Chiarawongse et al. (2012) inequalities among expected returns are employed to express the mentioned qualitative views. A Markov chain Monte Carlo simulation algorithm is
provided to solve the problem efficiently.
The optimization and performance of a sectorial cryptocurrency classification portfolio is identified and described in terms of potential benefits. By
testing six optimization targets, the resulting portfolios are compared through the CRIX index (which represents the crypto market). Based on sales
orders, this work addresses a portfolio optimization problem. In it, the investor must sell those risky assets whose prices have reached pre-established
sale thresholds. Since the portfolio problem is formulated as a sequential decision one, the investment strategy, which is actually a three-stage,
Čuljak et al (2022) automatic trading system, is revised at the beginning of each period. In the first stage, a historic data based, power fuzzy number is employed to
calculate the return of the investment strategy. This allows proposing a fuzzy mean, semi-variance portfolio optimization model, coupled to sales
orders. In stage 2, the proposed model is solved by designing a multi-objective genetic algorithm. In stage 3, an optimal investment strategy is selected
from a series of efficient solutions based on the Value-at-Risk ratio. Two real-market case studies are conducted to illustrate the practical application of
the model and algorithm proposed in this work.
The aim of this paper is to introduce a fuzzy, multi-attribute, group decision-making technique, considering the degrees of confidence of expert
opinions. The technique proceeds as follows:- A fuzzy decision matrix is built and normalized for each expert. - An integration or second-class
Debashree and Debajani aggregation matrix is built, condensing the individual opinion of each expert about an alternative, taking into account all its attributes. - For each
(2011) alternative, an agreement matrix among the n experts is built, using a fuzzy similarity measure. - An iterative process based on the fuzzy similarity
measures is conducted in order to reach consensus about all expert opinions.- Finally, and based on the consensual opinion of the group, a ranking
method is applied to the proposed alternatives.
This paper presents SINVLIO (Semantic Investment Portfolio), a tool based on semantic technologies and fuzzy logic techniques that recommends
García-Crespo et al., (2012) investments grounded in both psychological aspects of the investor and traditional financial parameters of the investments. The aim of the system
proposed in this paper is to combine the semantic approach with fuzzy logic techniques.
This work intends to provide the investor with more flexible decision possibilities by specifying risk tolerance and developing different investment
plans according to changing investment horizons. The resulting models are solved by a real-coded genetic algorithm. The models are tested by applying
Gupta et al (2021)
them to two real life instances. One of them involved 20 assets from the Indian National Stock Exchange. The second one dealt with 50 assets from the
S&P 500 and the NASDAQ-100 indexes.
An asset-return optimization problem is treated through fuzzy data analysis techniques. In order to optimize an investment portfolio, a series of non-
Hosseini and Hamidi (2016) linear mathematical models are applied, together with the “Change of Variables” technique. This allows the merging of two models and the creation of
integral, linear-model variables, which are worked out with a Lingo solver.
An optimization method based on the stochastic diffusion and stochastic differential equation is proposed for the constitution of a securities portfolio.
In researching the Shanghai Composite Stock Market Index, the authors obtained a new differential equation through dynamic programming. This
Huang (2020).
allowed obtaining the system’s optimal feedback control and optimal index. The automatic adjustment of the values of several parameter types allowed
obtaining the optimal securities portfolio. On these grounds, it was possible to provide more sensible stock price research and predictions.
Deep learning and natural language processing methods are applied to construct the view distribution in the Black-Litterman model. This approach is
Hung et al (2024) implemented to portfolio allocation. In this context, statistical analysis is performed in order to evaluate the performance of the different portfolios in
the short and long term. The benefits of this modeling approach are shown through an empirical analysis.
An investment-portfolio model focused on entropy is proposed. A multi-objective non-linear problem (MONLP) is solved through a fuzzy-programming
Jana et al. (2009)
approximation technique. The model intends to optimize a diversified asset portfolio, taking into consideration transaction costs and liquidity.
The portfolio optimization model that uses the L1 risk function (absolute mean risk deviation) is shown to be equivalent, under certain assumptions, to
Konno and Yamazaki (1991) the classical model of Markowitz. The linear nature of the implemented model overcomes the computational difficulties that arise in handling practical
problems associated to the classical model of Markowitz.
The stochastic dominance (SD) decision criterion and the empirical likelihood (EL) estimation method allowed the development of a new investment
Post et al (2018). portfolio optimization method. The desirable properties of large samples for the SD/EL solution portfolio were established, considering a particular
type of weakly dependent processes.
A cluster-based asset selection process is implemented, followed by the assessment of its advantages and disadvantages. Candidate assets are separated
Sass and Thos (2021) by none overlapping clusters, inside of which the assets are sorted according to their Sharpe ratio. This allows composing a new portfolio made up of
the best assets, all of them with equal weights.
A linear approach to qualitative portfolio analysis resorts to very specific heuristics sometimes carrying sufficient information to anticipate all the
qualitative rank of solutions. It is shown that correct analysis of optimal portfolios makes it necessary to apply complementary quantitative analysis.
Tarrazo (2016)
This is so because return to risk quality varies across efficient portfolios, whereas those in which variance is minimized may not be capable of risk
minimization.
This approach is based on a Fuzzy multi-criteria decision model (FMCDM) for investments in Information Technology/Information Systems (IT/IS),
supported on quantitative features expressing the opinion of the shareholders. The FMCDM comprises an evaluation process in two stages: criterion
weighing and alternative assessment. The evaluation process comprises the following steps:- Forming two evaluation groups: a weighing team and an
assessment team.- Determining the evaluation criteria.- Identifying possible alternative projects and combining them with the existing IT/IS portfolio.-
Establishing appropriate linguistic (quantitative) scales for the criteria.- Opinion of each member about the relative importance of the evaluation
Tzy-Yuan et al (2006)
criteria by means of pairwise comparison.The mathematical procedure comprises the following steps:- Fuzzy (triangular) criteria employed to start the
fuzzy reciprocal matrix of each member of the weighing team.- Calculating each local weight and the global weight.- Deciding the rules for the
translation of the assessment variables at a triangular fuzzy scale.- Calculating the final score of each alternative project based on the scores provided
by the members of the assessment team.- Translating each triangular fuzzy score given to the alternative projects into a neat value, in order to develop a
rank.

3
R.G. García-Cáceres et al. Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

opinions on the relative importance of each financial asset. through the following constraint:
Parameters: ∑N
Ajt: Difference between the average profitability of financial asset j j=1
α(ξ)j wj ≤ 1 (10)
and its profitability during period t, where (Ajt = rjt − rj .
The parameter that expresses the opinion of the experts about the
rj: Average expected profitability of financial asset j.
relative importance of the assets can be deterministic in nature when the
rjt: Profitability of asset j during period t.
DMs consensually agree. However, in public decision environments
ρ: Desired profit rate.
where this is not the case, it is usually stochastic. The probability dis­
vj: Percent bound of the capital invested in financial asset j.
tribution that describes the expert opinion on the success of the portfolio
T: Number of periods (t = 1,…,T).
can be previously calculated by means of techniques intended to
N: Number of financial assets (n = 1,., N).
determine the utility function of each DM or group of DMs (Keeney and
Variables:
Raiffa, 1976).
wj: Capital percentage invested in financial asset j.
∑ The utility function is the convex combination of two factors in j,
yt: Absolute value in the period t, where Nj=1 Ajt wj = yt
namely, parameter α(ξ)j , which corresponds to the stochastic opinion of
y: Optimal absolute value.
the experts about asset j, and variable wj , which expresses the capital
Objective:
fraction assigned to asset j. Just as the sum of all wj is 1, the sum of all
∑T
y α(ξ)j is also 1. Based on these inputs, the utility function, which is bound
y = Min t=1 t (1)
T within the (0, 1) range, represents a preference estimator of an invest­
Subjected to: ment portfolio’s utility.
The series of procedures applied at this stage allow obtaining the
- Absolute value linearization constraint: Joint Cardinal Probability Distribution of each alternative (f(l)), which
Investment portfolio profitability deviations can take any value expresses the odds of those optimal solutions with the same decision-
within real numbers because this is a condition of Ajt. The absolute variable configuration that governs the problem and corresponds to
value can be treated linearly with the aid of the following two the output variable resulting from the cardinal analysis.
constraints: The process starts by sorting the database that contains the simulated
share prices. Subsequently, each share’s quarterly profitability (rjt),
∑N
− Ajt wj + yt ≥ 0 t = {1, ….., T} (2) average profitability (rj) and profitability deviations (Ajt) are estimated.
In the current assessment, fixed term deposit (FTD) was chosen as the
j=1

∑N minimum accepted rate, since it represents the minimum safety bound,

− Ajt wj + yt ≤ 0 t = {1, ….., T} (3)
j=1 therefore corresponding to the lowest utility expected by the investor,
which was set at 2.36% for the four studied periods.
Expert opinions on investment portfolios could be provided by ex­
- Minimum revenue expectation constraint: perts or simulated through (0,1) random vectors for each α(ξ)j , thus
∑N
rj wj ≥ ρ (4) satisfying the condition established in Eq. 9. In this way, a different L_1
j=1
instance was generated with each random vector. The simulation of 100
instances was used to estimate the joint cardinal probability distribution
(f(l)), the mean total utility of the optimal investment portfolio and its
- Budget constraint:
standard deviation.
∑N
wj = 1 (5) The optimization model was programmed in GAMS win32 24.1.3
j=1
and executed in a PC with 32-byte Windows operative system, 1.73 GHZ
Intel T2080 processor, 45 GB hard disk and 1 GB RAM memory.
- Investment constraint:
3.2.1. In order to determine the variable resulting from the cardinal stage,
wj ≤ vj j = {1, ….., N} (6) the following procedure was followed

a) In order to facilitate the process of illustrating the cardinal analysis, a

- Non-negativity constraints: more efficient process than the purely random one was chosen. A
wj ≥ 0 j = {1, ….., N} (7) hundred scenarios (labelled s hereafter) were configured instead,
featuring different α(ξ)j values across assets (Table 1 of the
yt ≥ 0 t = {1, ….., T} (8) Appendix).
b) The frequencies of the alternatives (i.e., specific portfolios) were
Novel consideration: The present work introduces the modeling of a estimated from the set of optimal solutions of the mathematical
constraint on the estimation of the joint success probability of the programming model (Table A2 in Appendix A). As a result, 69 sim­
portfolio. As such, this bound could be included in any mathematical ulations were found to satisfy the minimum expected profitability,
programming model aimed at portfolio investment optimization. thus being considered to be valid for the analysis.
Parameters: c) The frequencies of the portfolios were adjusted so that the variable
α(ξ)j: Marginal, stochastic opinion of the DM’s on the success prob­ resulting from the cardinal analysis f(l) summed 1. For such purpose,
ability of financial asset j (whether it is profitable or not). the probability obtained from the simulation, which is expressed in
Being: terms of frequency/100, is adjusted by multiplying said value by
∑N (100/69).
j=1
α(ξ)j = 1 (9)

The utility function of the importance of the assets is introduced Table 2 characterizes the results of the cardinal stage, detailing each
group of simulations sharing the same configuration, together with the

4
R.G. García-Cáceres et al. Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

objective function values they attained, which were mostly different economic, strategic, organizational, sociological, marketing or ac­
from one another. Twenty-eight configurations (l) were identified as counting related.
noteworthy to be studied, as it can be seen in Table A2 of the Appendix, According to Babi´c-Hodovi´c et al. (2011), good reputation directly
together with the statistics obtained from the simulation: investment affects revenues by increasing share prices and investment profitability.
percentages corresponding to each set of alternatives (f(l)); expected Reputation is a dynamic process that evolves every time there is new
profitability value (E[y(l)]); and associated investment risk (σ [y(l)]). relevant information (e.g., share prices). In turn, these evolutionary
The analysis doesn’t show considerable differences between the shifts constitute important inputs for decision-making. According to
probabilities of occurrence of the different configurations, which range Fama (1970), investors must constantly update the information with
between 1 and 7, alternative 1 being the best qualified one. As to the which they feed their models, so that their expectations come closer to
expected profitability value during the studied period, alternatives 21 the reality of the market.
and 1 are best ranked. Finally, risk assessment shows that alternatives 5 The qualitative stage, which is supported on SMAA-O, takes into
and 1 scored the best variation coefficients (σ[y(l)] / E[y(l)]), not taking consideration the criterion weights that support each alternative. The
into account those cases with few simulated data. On the whole, alter­ favorability rankings and the criterion weight features, both resulting
native 1 shows high performance in the cardinal analysis. from the application of the technique, allow the DMs to carry out their
Execution times were found to be acceptable since several instances work in public decision environments. At this stage, the ordinal
took around 3 min, with an average of 1.4 min. This is consistent with acceptability index measures each alternative’s dominance probability,
the results reported by Konno and Yamazaki (1991). Regarding while the central weight vector is actually the centroid of those weights
computational power, the current results are also acceptable, taking into favoring an alternative in a particular ranking. In this context, the first
account the scalability of our CPU times. This is so despite the fact that favorability ranking is the only one within the scope of the present work.
we had much larger computational power than the mentioned authors The alternatives of the problem are the set of asset combinations that
and that our model includes additional constraints to those executed in reached the optimum-portfolio status during the cardinal stage. These
their work. alternatives, which result from the solution of the L_1 model, have to be
Ordinal analysis. ordinally qualified later on. According to García-Cáceres et al. (2009),
This stage is based on Stochastic Multicriteria Acceptability Analysis this process imposes the need to specify the concept of class as a set of
– Ordinal (SMAA-O) (Lahdelma et al., 2003), which, in the context of alternatives with identical ordinal utility values. This particular feature
IAM, only deals with qualitative criteria (García-Cáceres et al., 2009), of the alternatives’ ordinal criteria within a given class allows a more
since this is the purpose of this part of the analysis. The procedure in efficient use of the technique, as it reduces the number of options that
question resorts to Likert tables to turn the qualitative aspects into need to be processed during the execution of SMAA-O, finally resulting
ordinal variables (Albaum, 1997). in faster CPU times.
Criterion: Reputation of the decision alternative. Each alternative is ordinally evaluated. Provided that alternatives
The qualitative criterion employed to assess this feature is Business are made up of particular sets of assets, and that they contain the val­
Reputation of an investment alternative. This feature is associated to uations of the reputation of the companies they include, the assessment
aspects such as trust, solidness and business tradition, among other of each specific portfolio is carried out through Likert tables. Penalty
considerations. Although questioned for lacking a clear and consistent assignment takes into consideration the following aspects:
definition, this criterion is still considered to be very important, to the
point that brokers and investors usually support their decisions on it, – Each share belongs to a single company.
especially in the case of long-term ventures, which justifies its inclusion – The first fifteen assets shown in the columns of the appendix (assets
in the current integral optimization (Barnett et al., 2006). Following the 1–15) are highly reputed, while the other 16 (16− 31) are considered
work of this author, business reputation estimation is based on the not to have such good reputation.
following premise: “it corresponds to the collective judgment of a given
business on the part of the DMs, taking into account its evolution over The penalties of the companies making up each decision alternative
time and its financial, social and environmental impacts”. are then summed. In this context, three Likert table categories were
Different guilds in the business sector have commissioned studies established by the team of researchers.
aimed at clearing the concept of business reputation and its decision-
making implications in the ambit of investment portfolios. In their • Ordinal 1: corresponds to those alternatives containing only ranked
work The Reputational Landscape, Fombrun and Riel (1997) propose six companies.
aspects that influence business reputation and allow classifying it as

Table 2
Characterization of the results of the cardinal stage.
l s f (l) E[y (l)] σ[y (l)] l s f (l) E[y (l)] σ[y (l)]
1 32, 34, 49, 50, 52, 56, 77 7/69 0.04175 0.0055 15 82, 85 2/69 0.02141 0.0046
2 43, 45, 51, 57, 79, 100 6/69 0.02617 0.0035 16 22, 38 2/69 0.01384 0.0054
3 4, 5, 12, 13, 26 5/69 0.02738 0.0052 17 40, 73 2/69 0.04278 0.0054
4 67, 18, 21, 25, 30 5/69 0.02110 0.0052 18 7 1/69 0.00590 0
5 72, 89, 95, 99 4/69 0.03472 0.0044 19 3 1/69 0.00590 0
6 33, 39, 58, 70 4/69 0.02307 0.0042 20 42 1/69 0.01441 0
7 24, 59, 63, 78 4/69 0.02768 0.0045 21 1 1/69 0.05813 0
8 48, 69, 96 3/69 0.02304 0.0031 22 23 1/69 0.02634 0
9 55, 64, 66 3/69 0.02381 0.0043 23 68 1/69 0.02305 0
10 11, 27, 80 3/69 0.02147 0.0044 24 76 1/69 0.04068 0
11 9, 14 2/69 0.00590 0.0032 25 84 1/69 0.03704 0
12 28, 31 2/69 0.02075 0.0037 26 71 1/69 0.03702 0
13 17, 37 2/69 0.01423 0.0048 27 74 1/69 0.00590 0
14 41, 94 2/69 0.03615 0.0050 28 90 1/69 0.01848 0

Where: l: set of alternatives; f(l): adjusted marginal probability; E[y(l)]: expected profitability value; σ[y(l)]: profitability standard deviation. E[y(l)].

5
R.G. García-Cáceres et al. Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

• Ordinal 2: corresponds to those alternatives that include only one Table 4

outranked company. Ordinal output of the problem.
• Ordinal 3: corresponds to those alternatives including 2–16 out­ 11 1
ranked companies.
The remaining alternatives 0
• The resulting Likert table is shown in Table 3. These ordinal values
intend to establish a strong differentiation among categories.

. Table 5
Results of the integral analysis.
Table 3 presents the set of alternatives (l), which were chosen upon
the cardinal analysis, together with the assets making up each alterna­ l1 p1 (l1,11) dl1
tive and it’s associated ordinal score. This table is based on Table A2 of 11 3/69 1
the Appendix, which presents a likert scale detailing each of the 69 The remaining alternatives 0 0
evaluated scenarios. As an example, alternative 4 contains assets 16, 17
and 26, which are related to companies whose reputation is not
(0.55%), as indicated by its variation coefficient (0.132). This portfolio
considerably high. As a consequence, the ordinal value associated to this
contains several non-highly reputed assets, which makes it not adequate
alternative is 3.
for short-term Investments. Alternative to has the second highest
As it can be observed in the table above, the portfolios associated to
average profitability 2.617%, coupled to a 0.35% volatility and a vari­
28 alternatives cover the range of all possible values. Alternative 11 is
ation coefficient of 0.134, as identified during the evaluation period this
the best ranked one, with a high ordinal score (1). It is made up of four
portfolio has two assets with different reputation levels.
highly reputed assets: 2, 10, 11, 15, as shown by the ordinal accept­
The integral optimization of investment portfolios, conducted here
ability index (which is 1 for the first ranking (bl1 )), obtained from the with the support of IAM, allows the technical incorporation of expert
application of SMAA-O (Lahdelma and Salminen, 2003) (Table 4). opinions on the performance of a given asset in a particular time hori­
zon. This result, which is complementarily reached both from the car­
dinal and ordinal perspectives, constitutes an important step towards a
3.3. Integration analysis
more realistic solution of the problem. Although it was just an example,
the results of the cardinal and ordinal analyses showed important dif­
This last stage of IAM resorts to the deterministic SMAA technique
ferences, which can only be conciliated through the integration analysis.
introduced by Lahdelma et al. (2002). The inputs at this stage are the
These differences highlight both the main difficulties faced in this type
variables resulting from the previous two stages: ordinal acceptability
of endeavor and the usefulness of IAM to surmount them.
index (bl1 ) and joint cardinal probability distribution f(l).
( )
The resulting variables at this stage are joint integral indexes p1 el 4. Conclusions and research perspectives
and integral acceptability indexes dl1 . The former is a joint probability
function that results from multiplying the variables coming from the The purpose of the present research is to complement traditional
cardinal stage by those coming from the ordinal stage (f(l) * bl1), under portfolio analyses reported in the literature, by including novel aspects,
the independence assumption. In turn, the integral acceptability indexes especially human opinions. It is emphasized that the best solutions can
dl1 are obtained with the aid of the deterministic SMAA (Lahdelma and be reached by combining human and machine efforts. The objective of
Salminen, 2001), having as input the two variables mentioned above, this study is to contribute new elements to portfolio development re­
which are shown in Table 5: searchers and practitioners, so that they can conduct more realistic an­
The indicators allow establishing that alternative 11 have the best alyses as compared to traditional ones.
integral investment features (see the table of solutions to the instances New research perspectives point at: other qualitative specificities of
considered in Appendix A, Table A2). According to the first six alter­ investment portfolio problems such as cultural or political criteria;
natives, the average percentages to be invested in the shares of com­ exploring new ways of including (modeling) expert opinions in invest­
panies 2, 10, 11 and 15 are 28.2%, 4%, 46.9% and 20.92%, respectively. ment portfolio models and comparing the results to those obtained by
( )
Alternative 11 shows the highest joint integral index p1 el and integral related works; conducting case studies in companies dedicated to the
stocks market, in order to evaluate the efficacy of the current approach
acceptability index dl1 values in the studied ranking, which are respec­
to the problem. Also important is the need to study the solution pro­
tively (3/69) and 1, while the other alternatives exhibit a value of zero.
cedure of the model employed for the cardinal analysis, so as to improve
It is worthwhile noting the weaknesses of the portfolio represented
its computational performance.
by alternative 11, which showed a relatively low average profitability
(2.147%) during the four studied periods, counterbalanced by a lower
Funding acknowledgement
volatility (0.32%). In sum, these companies are strongly recommended
investment objectives under the current analysis.
This project does not contain financing.
In spite of the above, some other alternatives could also be consid­
ered by the investors, namely, numbers 1 and 2. The former one is the
most profitable one (4.175%), although it has a relatively high volatility

Table 3
Ordinal criterion.
l j Ordinal value l j Ordinal value l j Ordinal value l j Ordinal value

1 16, 17 3 2 5, 16 2 3 15, 16, 19 3 4 16, 17, 26 3

5 4, 16 2 6 15, 16 2 7 16, 17, 29 3 8 6, 16 2
9 8, 16 2 10 12, 16, 17 3 11 2, 10, 11, 15 1 12 2, 12, 16 2
13 2, 16, 26 3 14 2, 16 2 15 15, 16, 26 3 16 16, 20, 21 3
17 17, 21 3 18 1, 2, 23, 29 3 19 1, 15, 16, 29 3 20 15, 19, 20 3
21 1, 17, 20 3 22 2, 16, 20 3 23 2, 16, 23, 28 3 24 4, 9, 10, 16 2
25 4, 16, 17 3 26 5, 15, 16 2 27 5, 11, 12, 16 2 28 5, 16, 26 3

6
R.G. García-Cáceres et al. Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

Ethical statement consequences.

To verify originality, your article may be checked by the originality
Hereby, I Rafael Guillermo García Cáceres consciously assure that for detection software iThenticate. See.
the manuscript “An approach to. also http://www.elsevier.com/editors/plagdetect.
the integral optimization of investment portfolios” the following is I agree with the above statements and declare that this submission
fulfilled: follows the policies of Solid State.
1) This material is the authors’ own original work, which has not Ionics as outlined in the Guide for Authors and in the Ethical
been previously published elsewhere. Statement.
2) The paper is not currently being considered for publication
elsewhere. CRediT authorship contribution statement
3) The paper reflects the authors’ own research and analysis in a
truthful and complete manner. García-Cáceres Rafael Guillermo: Conceptualization, Methodol­
4) The paper properly credits the meaningful contributions of co- ogy, article writing. Franklin Ignacio Páez-Rivera: Methodology, Su­
authors and co-researchers. pervision. Bernarda Aldana-Gómez: Formal analysis,
5) The results are appropriately placed in the context of prior and Conceptualization. Ernesto Acosta-Gempeler: Data curation, article
existing research. writing. John Wilmer Escobar-Velásquez: Software, Validation.
6) All sources used are properly disclosed (correct citation). Literally
copying of text must be indicated as.
such by using quotation marks and giving proper reference. Declaration of Competing Interest
7) All authors have been personally and actively involved in sub­
stantial work leading to the paper, and. The authors declare that they have no known competing financial
will take public responsibility for its content. interests or personal relationships that could have appeared to influence
The violation of the Ethical Statement rules may result in severe the work reported in this paper.

Appendix A. Association (After Pareto analysis) of asset combinations, percentage to be invested in each share, and investment portfolio
risk and profitability
Table A1
Instances considered

OPINION ABOUT THE RELATIVE IMPORTANCE OF ASSET j

Combination. α1 α2 α3 α4 α5 α6 α7 α8 α9 α10 α11 α12 α13 α14 α15 α16 α17 α18 α19 α20 α21 α22 α23 α24 α25 α26 α27 α28 α29 α30 α31
1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
2 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
3 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
4 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
5 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
6 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
7 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
8 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
9 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
10 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
11 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
12 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
13 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
14 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
15 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
17 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
18 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
19 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
20 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
21 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0
22 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0
23 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
24 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
25 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0
26 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0
27 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0
28 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0
29 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
30 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
31 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
32
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
33 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 3/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
34 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 4/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
35 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
(continued on next page)

7
R.G. García-Cáceres et al. Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

Table A1 (continued )

1/ 5/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
36 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 6/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
37 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 7/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
38 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 8/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
39 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 9/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
40 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 10/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
41 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 11/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
42 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 12/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
43 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 13/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
44 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 14/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
45 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 15/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
46 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 16/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
47 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 17/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
48 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 18/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
49 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 19/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
50 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 20/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
51 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31
1/ 21/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
52 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31
1/ 22/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
53 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31
1/ 23/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
54 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31
1/ 24/ 1/ 1/ 1/ 1/ 1/ 1/
55 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31
1/ 25/ 1/ 1/ 1/ 1/ 1/
56 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31
1/ 26/ 1/ 1/ 1/ 1/
57 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31
1/ 27/ 1/ 1/ 1/
58 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31
1/ 28/ 1/ 1/
59 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31
1/ 29/ 1/
60 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31
1/ 30/
61 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 2/ 1/
62 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 3/ 1/
63 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 4/ 1/
64 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 5/ 1/
65 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 6/ 1/
66 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 7/ 1/
67 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 8/ 1/
68 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 9/ 1/
69 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 10/ 1/
70 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 11/ 1/
71 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 12/ 1/
72 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 13/ 1/
73 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
(continued on next page)

8
R.G. García-Cáceres et al. Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

Table A1 (continued )

1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 14/ 1/
74 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 15/ 1/
75 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 16/ 1/
76 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 17/ 1/
77 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 18/ 1/
78 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 19/ 1/
79 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 20/ 1/
80 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 21/ 1/
81 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 22/ 1/
82 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 23/ 1/
83 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 24/ 1/
84 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 25/ 1/
85 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 26/ 1/
86 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31 31
1/ 1/ 1/ 27/ 1/
87 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31 31
1/ 1/ 28/ 1/
88 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31 31
1/ 29/ 1/
89 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31 31
30/ 1/
90 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
31 31
2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
91 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
92 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
93 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
94 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
95 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
96 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
97 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
98 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
99 0 0
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31
1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 2/ 2/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/ 1/
100 0 0 31
31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31 31

9
 Table A2

R.G. García-Cáceres et al.

Solutions to the instances considered
Percentage to be invested in each share

s w1 w2 w3 w4 w5 w6 w7 w8 w9 w10 w11 w12 w13 w14 w15 w16 w17 w18 w19 w20 w21 w22 w23 w24 w25 w26 w27 w28 w29 w30 w31
1 44.0% 33.4% 22.6%
2 11.3% 7.5% 1.5% 79.7%
3 10.7% 40.4% 3.9% 45.1%
4 44.0% 33.4% 22.6%
5 44.0% 33.4% 22.6%
6 4.2% 55.6% 27.4% 12.7%
7 4.6% 18.4% 9.9% 67.1%
8 6.1% 42.9% 8.4% 42.7%
9 28.2% 4.0% 46.9% 20.9%
10 40.1% 14.2% 45.7%
11 42.1% 32.2% 25.7%
12 44.0% 33.4% 22.6%
13 44.0% 33.4% 22.6%
14 28.2% 4.0% 46.9% 20.9%
15 61.8% 20.1% 18.1%
16 53.2% 18.0% 24.2% 4.6%
17 45.2% 38.7% 16.1%
18 36.7% 34.9% 28.4%
19 54.7% 29.8% 2.6% 12.8%
20 25.1% 2.6% 59.0% 13.3%
21 36.7% 34.9% 28.4%
22 22.9% 31.7% 45.4%
23 44.1% 42.9% 13.1%
24 36.7% 34.9% 28.4%
25 36.7% 34.9% 28.4%
10

26 44.0% 33.4% 22.6%

27 42.1% 32.2% 25.7%
28 37.5% 27.0% 35.5%
29 8.2% 16.1% 20.2% 55.5%

Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

30 36.7% 34.9% 28.4%
31 37.5% 27.0% 35.5%
32 79.2% 20.8%
33 35.4% 64.7%
34 79.2% 20.8%
35 3.8% 22.0% 74.2%
36 51.1% 46.7% 2.3%
37 3.1% 64.1% 32.8%
38 36.9% 42.8% 20.4%
39 35.4% 64.7%
40 70.5% 29.5%
41 23.9% 76.1%
42 17.4% 43.8% 38.8%
43 84.3% 15.7%
44 9.5% 0.1% 0.8% 52.5% 37.1%
45 42.8% 57.2%
46 49.5% 3.5% 38.5% 8.6%
47 15.8% 36.3% 48.0%
48 45.9% 54.1%
49 79.0% 21.0%
50 79.0% 21.0%
51 57.6% 42.4%
52 79.0% 21.0%
53 11.4% 71.5% 17.1%
54 22.9% 76.1% 1.0%
(continued on next page)
 R.G. García-Cáceres et al.
Table A2 (continued )
55 26.7% 73.3%
56 79.0% 21.0%
57 57.6% 42.4%
58 28.1% 71.9%
59 83.0% 15.6% 1.4%
60 59.3% 35.3% 1.9% 3.5%
61 0.2% 22.5% 67.7% 9.6%
62 39.6% 6.6% 53.8%
63 37.3% 2.1% 60.6%
64 33.9% 66.1%
65 71.4% 12.3% 16.4%
66 33.9% 66.1%
67 56.3% 2.6% 41.1%
68 15.5% 65.8% 2.3% 16.4%
69 60.6% 39.4%
70 35.4% 64.7%
71 23.5% 22.5% 54.0%
72 59.8% 40.2%
73 73.5% 26.5%
74 14.4% 13.3% 30.3% 42.1%
75 6.7% 16.8% 59.1% 16.3% 1.1%
76 11.9% 0.4% 22.5% 65.2%
77 79.0% 21.0%
78 76.4% 12.1% 11.6%
11

79 43.4% 56.6%
80 17.2% 74.6% 8.3%
81 12.9% 57.5% 29.6%
82 21.0% 72.1% 6.9%

Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

83 64.2% 11.8% 24.1%
84 23.2% 70.2% 6.6%
85 21.0% 72.1% 6.9%
86 13.1% 59.6% 27.3%
87 25.7% 60.6% 13.7%
88 29.5% 8.7% 61.8%
89 41.9% 58.1%
90 29.3% 57.3% 13.4%
91 37.4% 62.6%
92 3.1% 62.1% 34.8%
93 64.4% 19.2% 16.4%
94 26.3% 73.7%
95 59.8% 40.2%
96 60.6% 39.4%
97 23.7% 0.4% 75.9%
98 8.1% 91.9%
99 59.8% 40.2%
100 51.2% 48.8%
R.G. García-Cáceres et al. Journal of Open Innovation: Technology, Market, and Complexity 10 (2024) 100235

References Hosseini, S., Hamidi, N., 2016. Common Funds Investment Portfolio Optimization with
Fuzzy Approach. Procedia Econ. Financ. 36, 96–107. https://doi.org/10.1016/
s2212-5671(16)30020-x.
Albaum, G., 1997. The Likert Scale Revisited. J. Mark. Res. Soc. 39 (2), 331–348.
Huang, Y., 2020. Portfolio optimization based on jump-diffusion stochastic differential
Babi´c-Hodovi´c, V., Mehi´c, E., Arslanagi´c, M., 2011. Influence of Banks’ Corporate
equation. Alex. Eng. J. (4), 2503–2512.
Reputation on Organizational Buyers Perceived Value. Procedia Soc. Behav. Sci. 24,
Hung, M.C., Hsia, P.H., Kuang, X.J., Lin, S.K., 2024. Intelligent portfolio construction via
351–360.
news sentiment analysis. Int. Rev. Econ. Financ., 89, 605–617.
Barbati, M., Figueira, J.R., Greco, S., Ishizaka, A., Panaro, S., 2023. A multiple criteria
Jana, P., Roy, T., Mazumder, S., 2009. Multi-objective possibilistic model for portfolio
methodology for priority based portfolio selection. Socio-Econ. Plan. Sci., 101595
selection with transaction cost. J. Comput. Appl. Math. 228, 188–196.
Barnett, M., Jermier, J., Lafferty, B., 2006. Corporate Reputation: The Definitional
Kandakoglu, M., Walther, G., Ben Amor, S., 2023. The use of multi-criteria decision-
Landscape. In: Corporate Reputation Review, 9. Palgrave – Journals,, pp. 26–38.
making methods in project portfolio selection: a literature review and future
Celikyurt, U.&, Özekici, Z., 2007. Multiperiod portfolio optimization models in stochastic
research directions. Ann. Oper. Res. 1–24.
markets using the mean–variance approach. Eur. J. Oper. Res. 179 (1), 189–202.
Keeney, R.L., Raiffa, H., 1976. Decisions with Multiple Objectives: Preferences and Value
Chen, L.H., Huang, L., 2009. Portfolio optimization of equity mutual funds with fuzzy
Tradeoffs. Wiley,, New York.
return rates and risks. Expert Syst. Appl. 36 (2), 3720–3727.
Konno, H., Yamazaki, H., 1991. Mean – Absolute deviation portfolio optimization model
Chiarawongse, A., Kiatsupaibul, S., Tirapat, S., Van Roy, B., 2012. Portfolio selection
and its applications to Tokyo stock market. Manag. Sci. 37, 519–531.
with qualitative input. J. Bank. Financ. 36 (2), 489–496.
Lahdelma, R., Salminen, P., 2001. SMAA-2: Stochastic multicriteria acceptability analysis
Clyman, D., 1995. Unreasonable Rationality? Manag. Sci. 41, 1538–1548.
for group decision making. Oper. Res. 49 (3), 444–454.
Čuljak, M., Tomić, B., Žiković, S., 2022. Benefits of sectoral cryptocurrency portfolio
Lahdelma, R., Miettinen, K., Salminen, P., 2003. Ordinal criteria in stochastic
optimization. Res. Int. Bus. Financ., 101615
multicriteria acceptability analysis (SMAA). Eur. J. Oper. Res. 147 (1), 117–127.
Debashree, G., Debajani, C., 2011. Fuzzy multi attribute group decision making method
Markowitz, H., 1952. Portfolio selection. J. Financ. 7 (1), 77–91.
to achieve consensus under the consideration of degrees of confidence of experts’
Post, T., Karabatı, S., Arvanitis, S., 2018. Portfolio optimization based on stochastic
opinions. Comput. Ind. Eng. 60, 493–504.
dominance and empirical likelihood. J. Econ. 206 (1), 167–186.
Fama, E., 1970. Efficient capital markets: A review of theory and empirical work.
Sass, J., Thös, A., 2024. Risk reduction and portfolio optimization using clustering
J. Financ. vol. XXV (n◦ 2), 383–417.
methods. Econ. Stat. (in press).).
Fombrun, C., Riel, C.B.M., 1997. The Reputational Landscape. Corp. Reput. Rev. 1 (1),
Sharpe, S.A. 1999. Stock prices, expected returns, and inflation. Finance and Economics
5–13.
Discussion. Series 1999-02, Board of Governors of the Federal Reserve System (U.S.).
García-Cáceres, R.G., Aráoz, J., Palacios, F., 2009. Integral analysis method – IAM. Eur. J.
Tarrazo, M., 2016. Portfolio heuristics, linearity, and qualitative analysis. Appl. Econ.
Oper. Res. 192, 891–903.
Financ. 3 (4), 134-147.
García-Cáceres, R.G., Vega-Mejía, C., Caballero-Villalobos, J.P., 2011. In: Integral
Tzy-Yuan, C., Seng-Cho, C., Gwo-Shiug, T., 2006. Evaluating IT/IS Investments: A fuzzy
Optimization of the Container Loading Problem. In “Stochastic Optimization, Seeing
multi-criteria decision model approach. Eur. J. Oper. Res. 173, 1026–1046.
the optimal for the uncertain”. InTech,, Rijeka (Croatia), pp. 225–254.
Xie, D., 2021. Empirical Study of Markovitz Portfolio Theory and Model in the Selection
García-Crespo, A., López, J., González, I., Colomo, R., Ruiz, B., 2012. SINVLIO: Using
of Optimal Portfolio in Shanghai Stock Exchange of China. J. Econ., Bus. Manag. 9
semantics and fuzzy logic to provide individual investment portfolio
(4), 87–92.
recommendations. Knowl. – Based Syst. 27, 103–118.
Gupta, P., Mehlawat, M.K., Khan, A.Z., 2021. Multi-period portfolio optimization using
coherent fuzzy numbers in a credibilistic environment. Expert Syst. Appl. 114135.

12