POSSIBILISTIC SHARPE RATIO BASED NOVICE PORTFOLIO SELECTION MODELS

Rupak Bhattacharyya et al. (Eds) : ACER 2013,
pp. 33–45, 2013. © CS & IT-CSCP 2013 DOI : 10.5121/csit.2013.3204
POSSIBILISTIC SHARPE RATIO BASED NOVICE
PORTFOLIO SELECTION MODELS
Rupak Bhattacharyya
Department of Applied Science & Humanities, Global Institute of Management
& Technology, Krishnagar- 741102, West Bengal, India
mathsrup@gmail.com
ABSTRACT
This paper uses the concept of possibilistic risk aversion to propose a new approach for
portfolio selection in fuzzy environment. Using possibility theory, the possibilistic mean,
variance, standard deviation and risk premium of a fuzzy number are established. Possibilistic
Sharpe ratio is defined as the ratio of possibilistic risk premium and possibilistic standard
deviation of a portfolio. The Sharpe ratio is a measure of the performance of the portfolio
compared to the risk taken. The higher the Sharpe ratio, the better the performance of the
portfolio is and the greater the profits of taking risk. New models of fuzzy portfolio selection
considering the possibilistic Sharpe ratio, return and skewness of the portfolio are considered.
The feasibility and effectiveness of the proposed method is illustrated by numerical example
extracted from Bombay Stock Exchange (BSE), India and is solved by multiple objective genetic
algorithm (MOGA).
KEYWORDS
Fuzzy portfolio selection problem, Possibilistic Sharpe ratio, Possibilistic risk premium, Utility
function, Multiple objective genetic algorithm.
1. INTRODUCTION
The basic objective of portfolio selection problem is the optimal allocation of money in different
stocks so that it maximizes the return and minimizes the risk of investment. Markowitz [1] who
integrates probability theory and optimization to model the problem does the first mathematical
formulation of portfolio selection problem in this return-risk framework. The mean-variance
model given by him is valid only when the returns of the stocks are normally distributed i.e.,
symmetric; which is not true in most of the cases. However, this model represents the risk adverse
nature of the investors. To deal with the asymmetric nature of the return, skewness has been
incorporated in the model by many researchers like Lai [2], Konno and Shirakawa [3], Konno and
Suzuki [4], Chunhachinda et al. [5], Liu et al. [6], Prakash et al. [7], Briec et al. [8], Yu et al. [9],
Bhattacharyya et al. ([10], [11], [12]), Chatterjee et al. [13] and others. Consideration of variance
as risk is erroneous as it equally suggests penalties for up and down deviations from the mean. To
face this problem, Markowitz [14] recommends semi-variance, a downside risk measure. Another
alternative definition of risk is the probability of an adverse outcome (Roy [15]). The popular risk
measure Value at Risk (Castellacci and Siclari [16], Philippe [17]) is in fact an alternative
expression of the definition by Roy [15]. Different authors like Philippatos and Wilson [18],
Philippatos and Gressis [19], Nawrocki and Harding [20], Simonelli [21], Huang [22], Qin et al.
[23], Bhattacharyya et al. [10] use entropy as an alternative measure of risk to replace variance
proposed by Markowitz [1].

34 Computer Science & Information Technology (CS & IT)
Since the Sharpe ratio has been derived in 1966 by William Sharpe [24], it has been one of the
most referred risk/return measures used in finance, and much of this popularity can be attributed
to its simplicity. The ratio's credibility has been boosted further when Professor Sharpe won a
Nobel Memorial Prize in Economic Sciences in 1990. The Sharpe ratio is defined as the ratio
between the risk premium and the standard deviation. It is a risk-adjusted measure of return that is
often used to evaluate the performance of a portfolio. The ratio helps to make the performance of
one portfolio comparable to that of another portfolio by making an adjustment for risk. The idea
of the ratio is to see how much additional return you are receiving for the additional volatility of
holding the risky asset over a risk-free asset - the higher the better. However, the problem with
Sharpe ratio is the presence of standard deviation in the formula. The issue with this formula lies
in its application to investments or securities that do not have normally distributed returns.
Nevertheless, consideration of positively skewed returns can solve this problem.
In most of the research works on portfolio selection, the common assumptions are that the
investor have enough historical data and that the situation of asset markets in future can be
reflected with certainty by asset data in past. However, it cannot always be made with certainty.
The usual feature of financial environment is uncertainty. Mostly, it is realized as risk uncertainty
and is modelled by stochastic approaches. However, the term uncertainty has the second aspect-
vagueness (imprecision or ambiguity) which can be modelled by fuzzy methodology. In this
respect, to tackle the uncertainty in financial market, fuzzy, stochastic-fuzzy and fuzzy-stochastic
methodologies are extensively used in portfolio modelling. By incurring fuzzy approaches
quantitative analysis, qualitative analysis, experts’ knowledge and investors’ subjective opinions
can be better integrated into a portfolio selection model. Authors like Konno and Suzuki [25],
Leon et al. [26], Vercher [27], Bhattacharyya et al. [12] and others use fuzzy numbers to embody
uncertain returns of the securities and they define the portfolio selection as a mathematical
programming problem in order to select the best alternative. In possibilistic portfolio selection
models, two types of approaches are noticed. The return of a security is considered either as a
possibilistic variable or as a fuzzy number. In the later case, the possibilistic moments of the
fuzzy numbers are considered. Possibilistic portfolio models integrate the past security data and
experts’ judgment to catch variations of stock markets more plausibly. Tanaka and Guo [28]
propose two kinds of portfolio selection models by utilizing fuzzy probabilities and exponential
possibility distributions, respectively. Inuiguchi and Tanino [29] introduce a possibilistic
programming approach to the portfolio selection problem under the minimax regret criterion. Lai
et al. [30], Wang and Zhu [31] and Giove et al. [32] construct interval-programming models for
portfolio selection. Ida [33] investigates portfolio selection problem with interval and fuzzy
coefficients, two kinds of efficient solutions are introduced: possibly efficient solution as an
optimistic solution, necessity efficient solution as a pessimistic solution. Carlsson et al. [34]
introduce a possibilistic approach for selecting portfolios with the highest utility value under the
assumption that the returns of assets are trapezoidal fuzzy numbers. Fang et al. [35] propose a
portfolio-rebalancing model with transaction costs based on fuzzy decision theory. Wang et al.
[36] and Zhang and Wang [37] discuss the general weighted possibilistic portfolio selection
problems. Moreover, Lacagnina and Pecorella [38] develop a multistage stochastic soft
constraints fuzzy program with recourse in order to capture both uncertainty and imprecision as
well as to solve a portfolio management problem. Lin et al. [39] propose a systematic approach
by incorporating fuzzy set theory in conjunction with portfolio matrices to assist managers in
reaching a better understanding of the overall competitiveness of their business portfolios. Huang
[40] presents two portfolio selection models with fuzzy returns by criteria of chance represented
by credibility measure. Fei [41] studies the optimal consumption and portfolio choice with
ambiguity and anticipation. Zhang et al. [42] assume that the rates of return of assets can be
expressed by possibility distribution. They propose two types of portfolio selection models based
on upper and lower possibilistic means and possibilistic variances and introduce the notions of
lower and upper possibilistic efficient portfolios. Li and Xu [43] deal with a possibilistic portfolio
selection problem with interval center values. Parra et al. [44] introduce vague goals for return

Computer Science & Information Technology (CS & IT) 35
rate, risk and liquidity based on expected intervals. Terol et al. [45] formulate a fuzzy
compromise programming to the mean-variance portfolio selection problem. Huang [46] proposes
a mean-semivariance model for describing the asymmetry of fuzzy returns. Huang [47] extends
the risk definition of variance and chance to a random fuzzy environment and formulates
optimization models where security returns are fuzzy random variables.
In this paper, we have defined the possibilistic Sharpe ratio as the ratio between the possibilistic
risk premium and possibilistic standard deviation. First, we have considered a bi-objective
optimize problem that maximizes the possibilistic Sharpe ratio as well as the possibilistic
skewness of the portfolio. As the Sharpe ratio prefers symmetric distribution, we maximize the
skewness to get rid of this problem and to give more preference to positively skewed returns. We
also have proposed six more models that represent different scenarios.
The construction of the paper is as follows. In section 2, the possibilistic mean, standard
deviation, skewness and risk premium of fuzzy numbers are derived. The outcomes are used to
define the possibilistic Sharpe ratio. In section 3, the portfolio selection models are modelled. In
section 4, a multiple objective genetic algorithm is discussed. In section 5, an example is provided
to illustrate the feasibility and effectiveness of the proposed model using stock price data in the
form of triangular fuzzy numbers extracted from Bombay Stock exchange (BSE). Finally, in
section 6, some conclusions are specified.
2. POSSIBILISTIC MOMENTS AND POSSIBILISTIC RISK AVERSION
Possibility theory is introduced by Zadeh [48] as an alternative to probability theory in the
treatment of uncertainty. Fuzzy numbers represent a significant class of possibility distribution.
The operations with fuzzy numbers can be done by Zadeh’s extension principle. Different authors
have dealt with the notion of possibilistic moments of fuzzy numbers in different times. Dubois
and Prade [49] propose the concept of interval-valued expectation of fuzzy numbers. Calsson and
Fuller [50] define the possibilistic mean and variance of a fuzzy number where Fuller and
Majlender [51] define the weighted possibilistic mean and variance of a fuzzy number. Liu and
Liu [52] propose a definition of expected value of a fuzzy variable based on the possibilistic
notion of credibility measure. Saeidifar and Pasha [53] define possibilistic moments of fuzzy
numbers. Georgescu [54] proposes risk aversion by possibility theory and find out the risk
premium by defining a new notion of possibilistic variance.
Definition 2.1 [54] Let A% be a fuzzy number. Also let its α-level set 1 2[ A] [a ( ),a ( )]=α
α α% be
such that 1 2a ( ) a ( ).≠α α The central value of [ A]α% is defined as
a ( )2
2 1a ( )12 1
1 1
C([ A] ) xdx ( a ( ) a ( )).
a ( ) a ( ) 2
= = +
− ∫
αα
α
α α
α α
%
Definition 2.2 [50] The expected value of a fuzzy number A% is defined by
1
2 10
E( A) [ a ( ) a ( )]d .= +∫ α α α α%
Definition 2.3 [50] The variance of a fuzzy number A% is defined by

1 1
2 2
1 1 2 2
0 0
1
2 2
1 2
0
Var( A) POS[A a ( )][a ( ) E( A)] d POS[A a ( )][a ( ) E( A)] d
= ([a ( ) E( A)] [a ( ) E( A)] )d
α α α α α α
α α α α
= ≤ − + ≥ −
− + −
∫ ∫
∫
% % % % %
% %
Definition 2.4 The skewness of a fuzzy number A% is defined by
3
3
M ( A)
Skew( A) ,
( Var( A))
=
%
%
%
where,
1 1
3 3
3 1 1 2 2
0 0
1
3 3
1 2
0
M ( A) POS[ A a ( )][a ( ) E( A)] d POS[ A a ( )][a ( ) E( A)] d
= ([a ( ) E( A)] [a ( ) E( A)] )d .
= ≤ − + ≥ −
− + −
∫ ∫
∫
α α α α α α
α α α α
% % % % %
% %
Definition 2.5 [54] The possibilistic risk premium A A, u=ρ ρ% % associated with the fuzzy
number A% and the utility function u is defined by,
Au( E( A) ) E(u( A)).− =ρ %
% %
Let us assume that the utility function u is twice differentiable, strictly concave and increasing.
Then the possibilistic risk premium A
ρ % has the form
A
1 u ( E( A))
V( A) ,
2 u ( E( A))
′′
≈ −
′
ρ %
%
%
%
where,
a ( )1 2
2
2 10 a ( )1
1
2 2
2 1 2 1
0
1
2 2
2 2 1 1 2 1
0
1
V( A) 2 ( x E( A)) dx d
a ( ) a ( )
2
= (a ( ) E( A)) (a ( ) E( A)) (a ( ) E( A))(a ( ) E( A)) d
3
2
= (a ( ) a ( )a ( ) a ( )) d 3E( A) (a ( ) a ( )) d
3
α
α
α α
α α
α α α α α α
α α α α α α α α α
 
 = −
− 
 
 − + − + − −
 
+ + − +
∫ ∫
∫
∫
% %
% % % %
%
1 1
2
0 0
1
2 2 2
2 2 1 1
0
3E ( A) d
2
= (a ( ) a ( )a ( ) a ( )) d E ( A).
3
α α α
α α α α α α
  
+ 
  
+ + −
∫ ∫
∫
%
%
Depending on different forms of the utility function, the risk premium will have different forms.

Exponential utility of the form ( ) x
u x 1 e= − κ
is unique in exhibiting constant absolute risk
aversion (CARA) and has been used successfully in portfolio selection problem. In this literature,
we consider the utility function as u(x) = 1 − e− 2x
. Then we have,
2x
2x
u ( E( A)) 4e
2;
u ( E( A)) 2e
−
−
′′ −
= = −
′
%
%
so that, A V( A).≈ρ %
%
Definition 2.6 We define the possibilistic Sharpe ratio (PSR) of a fuzzy number A% as,
A V( A)
PSR( A) .
Var( A) Var( A)
= =
ρ % %
%
% %
Example 2.7 Let A ( a,b,c )=% be a triangular fuzzy number. Then we have,
2 2 2
3 3 3 2 2 2
2 2 2 3
2 2 2
A
2 2 2
2 2
a 4b c
E( A) ,
6
1
Var( A) [ a +b +c ab bc ca],
18
[19(a c ) 8b 42b(a c ) (12b 15ac)(a c) 60abc]
Skew(A) ,
10 2( a +b +c ab bc ca)
1
[a +2b +c 2ab 2bc],
36
a +2b +c 2ab 2bc
PSR
6 2 a +b +c
+ +
=
= − − −
+ − − + + − + +
=
− − −
= − −
− −
=
ρ%
%
%
%
2
.
ab bc ca− − −
The results are obtained by definitions 2.2, 2.3, 2.4, 2.5 and 2.6.
3. POSSIBILISTIC PORTFOLIO SECTION MODEL
Let for i = 1, 2, …, n,
xi = the portion of the total capital invested in security i;
ip =% fuzzy number representing the closing price of the ith
security at present;
'
ip =% fuzzy number representing the estimated closing price of the ith
security for the next year;
di = fuzzy number representing the estimated dividend of the ith
security for the next year;
ir =% fuzzy number representing the return of the ith
security
'
i i i
i
p + d p
= .
p
−
As per discussion in the Introduction, we propose the following portfolio selection model in fuzzy
environment.

1 1 2 2 n n
1 1 2 2 n n
1 1 2 2 n n
n
i
i 1
i
Maximize PSR[r x r x .... r x ]
Maximize Skew[r x r x .... r x ]
subject to
E[r x r x .... r x ]( 3.1)
x 1
x 0,i 1,2,...,n.
α
=
+ + +

+ + +

 + + + ≥

 =


≥ =
∑
% % %
% % %
% % %
The constraint 1 1 2 2 n nE[r x r x .... r x ] α+ + + ≥% % % ensures that the expected return of the portfolio is
no less than a minimum desired value α. The second constraint
n
i
i 1
x 1
=
 
=  
 
∑ is the well-known
capital budget constraint on the assets. The last constraint [ ]ix 0 i≥ ∀ ensures that no short selling
is allowed in the portfolio here.
Note: The following models (3.2), (3.3), (3.4), (3.5), (3.6), (3.7) can also be considered by the
investors depending on their priorities.
1 1 2 2 n n
1 1 2 2 n n
1 1 2 2 n n
n
i
i 1
i
Maximize E[r x r x .... r x ]
subject to
Skew[r x r x .... r x ](3.2)
x 1
x 0,i 1,2,...,n.
β
=
+ + +

+ + +

 + + + ≥

 =


≥ =
∑
% % %
% % %
% % %
1 1 2 2 n n
1 1 2 2 n n
1 1 2 2 n n
n
i
i 1
i
subject to
E[r x r x .... r x ]
PSR[r x r x .... r x ]( 3.3 )
x 1
x 0,i 1,2,...,n.
α
γ
=
+ + +


 + + + ≥
 + + + ≥

 =


≥ =
∑
% % %
% % %
% % %
1 1 2 2 n n
1 1 2 2 n n
1 1 2 2 n n
n
i
i 1
i
subject to
E[r x r x .... r x ]
Skew[r x r x .... r x ](3.4 )
x 1
x 0,i 1,2,...,n.
α
β
=
+ + +


 + + + ≥
 + + + ≥

 =


≥ =
∑
% % %
% % %
% % %

1 1 2 2 n n
1 1 2 2 n n
1 1 2 2 n n
n
i
i 1
i
subject to
PSR[r x r x .... r x ]
Skew[r x r x .... r x ]( 3.5 )
x 1
x 0,i 1,2,...,n.
γ
β
=
+ + +


 + + + ≥
 + + + ≥

 =


≥ =
∑
% % %
% % %
% % %
and
1 1 2 2 n n
1 1 2 2 n n
1 1 2 2 n n
n
i
i 1
i
subject to( 3.6 )
x 1
x 0,i 1,2,...,n.
=
+ + +

+ + +
 + + +



 =


≥ =
∑
% % %
% % %
% % %
where the values of α, β, γ would be specified by the investors according to their needs.
Theorem 3.1 Suppose i i i ir = (a , b , c ), [i=1,2,...,n]% are independent triangular fuzzy numbers.
Then the model (3.1) generates the multi-objective programming problem model (3.8).
2 2 2n n n n n
i i i i i i i i i i
i = 1 i = 1 i = 1 i = 1 i = 1
n n n
i i i i i i
i = 1 i = 1 i = 1
1
M aximize a x + b x + c x a x b x
6 2
b x c x c x
( 3.7 )
         
−                  
         
    
− −       
    
∑ ∑ ∑ ∑ ∑
∑ ∑ ∑
1
2 2
n n n2
i i i i i i
i = 1 i = 1 i = 1
2n n n n n
i i i i i i i i i i
i = 1 i = 1 i = 1 i = 1 i = 1
a x a x + 2 b x
+ c x 2 a x b x 2 b x c x
−        
              
        
        
− −              
        
∑ ∑ ∑
∑ ∑ ∑ ∑ ∑
2 2 2
n n n n n
i i i i i i i i i i
i = 1 i = 1 i = 1 i = 1 i = 1
n n n
i i i i i i
i = 1 i = 1 i = 1
1
M aximize a x + b x + c x a x b x
10 2
b x c x c x

  

         
−                  
         
   
− −      
   
∑ ∑ ∑ ∑ ∑
∑ ∑ ∑
3
3 3n n n2
i i i i i i
i = 1 i = 1 i = 1
3 2n n n n
i i i i i i i i
i = 1 i = 1 i = 1 i = 1
a x 19 a x c x
8 b x 42 b x a x c x
−             +                           
       
− − +            
       
∑ ∑ ∑
∑ ∑ ∑ ∑
2
2
n n n n n
i i i i i i i i i i
i = 1 i = 1 i = 1 i = 1 i = 1
n n
i i i i
i = 1 i = 1
12 b x 15 a x c x a x c x
+ 60 a x b x
 
 
+ 
  
               
− +                     
             
 
  
 
∑ ∑ ∑ ∑ ∑
∑ ∑
n
i i
i = 1
n n n n
i i i i i i i i
i = 1 i = 1 i = 1 i 1
c x
subject to
1
a x + 4 b x + c x , x 1, x 0 ,i 1,2 ,...,n.
6
α
=
































   
       
   


  
≥ = ≥ =     
∑
∑ ∑ ∑ ∑

Proof: Since i i i ir = (a , b , c )% are triangular fuzzy numbers, by extension Principle of Zadeh it
follows that
n n n
1 1 2 2 n n i i i i i i
i = 1 i = 1 i = 1
r x r x .... r x = a x , b x , c x ,
 
+ + +   
 
∑ ∑ ∑% % % which is also a fuzzy
number. Combining this with the results obtained in example 2.7, we are with the theorem.
4. MULTIPLE OBJECTIVE GENETIC ALGORITHM
The proposed portfolio selection model (3.7) is solved by using Multiple Objective Genetic
Algorithm (MOGA). The MOGA proposed by Bhattacharyya et al. [55] is followed.
The following are followed for the development of the MOGA for the proposed model (3.7).
Representation: An n-dimensional real vector X = {x1, x2, …, xn} is used to represent a solution
where each ix [0,1], i =1, 2,..., n.∈
Initialization: L such solutions X1, X2, …, XL are randomly generated such that each of them
satisfies the constraints of the model. This solution set is the set P.
Cross Over and Mutation: Crossover operator is mainly responsible for the search of new strings.
Crossover operates on two parent solutions at a time and generates offspring solutions by
recombining both parent solution features. After selection of chromosomes for new population,
the crossover operator is applied. Here the arithmetic cross over is used. Mutation is the unary
operation by which genes present in a chromosome are changed. Here the usual mutation
procedure is followed.
Proposed multi-objective genetic algorithm has the following two important components:
(a) Consider a population P of feasible solutions of (3.7) of size L. We like to partition P into
subsets F1, F2, ..., Fk, such that every subset contain non-dominated solutions, but every solutions
of Fi are not dominated by any solution of Fi+1, for i = 1,2,..., k-1. Let the number of solutions of
P which dominate x is nx and the set of solutions of P that are dominated by x is Sx. Note that, as
there are two objective functions, these require O(2L2
) computations.
(b) To determine the distance of a solution from other solutions of a subset first sort the
subset according to each objective function values in ascending order of magnitude. For both
objective functions, the boundary solutions are assigned an inﬁnite distance value (a large value).
All other intermediate solutions are assigned a distance value for the objective, equal to the
absolute normalized difference in the objective values of two adjacent solutions. The overall
distance of a solution from others is calculated as the sum of individual distance values
corresponding to each objective.
For detailed discussions on ‘the division of P(T) into disjoint subsets having non-dominated
solutions’ and ‘distance of a solution of subset F from other solutions’, Roy et al. [56] can be
consulted.
Since two independent sorting of at most L solutions (in case the subset contains all the solutions
of the population) are involved, the above algorithm has O(2Llog L) computational complexity.

Using the above two operations proposed multi-objective genetic algorithm is formulated as:
1. Set probability of crossover Pc and probability of mutation Pm.
2. Set iteration counter T =1.
3. Generate initial population set of solution P(T) of size L.
4. Select solution P(T) for crossover and mutation.
5. Made crossover and mutation on selected solution and get the child set C(T).
6. Set 1P P(T ) C(T ).= U
7. Divide P1 into disjoint subsets having non-dominated solutions. Let these sets be F1, F2, ..., Fk.
8. Let 2 1 2 nP F F ... F .= U U U Select maximum integer L such that 2O( P ) L.≤
9. If O(P2) < L sort solutions of Fn+1 in descending order of their distance from other solutions of
the subset. Then select first L - O(P2) solutions from Fn+1 and add with P2.
10. Set T = T + 1 and P(T) = P2.
11. Go to step-4 if termination condition does not hold.
12. Output: P(T)
13. End algorithm.
Since in the above algorithm computational complexity of step-7 is O(2L2
), step-9 is O(2NlogN)
and other steps are ≤ O(N), so overall time complexity of the algorithm is O(2N2
).
In this MOGA, selection of new population after crossover and mutation on old population is
done by creating a mating pool by combining the parent and oﬀspring population and among
them best L solutions are taken as solutions of new population. By this way, elitism is introduced
in the algorithm.
5. CASE STUDY: BOMBAY STOCK EXCHANGE
In this section we apply our portfolio selection model on the data set extracted from Bombay
stock exchange (BSE). Bombay Stock Exchange is the oldest stock exchange in Asia with a rich
heritage of over 133 years of existence. What is now popularly known as BSE was established as
"The Native Share & Stock Brokers' Association" in 1875. It is the first stock exchange in India
which obtained permanent recognition (in 1956) from the Government of India under the
Securities Contracts (Regulation) Act (SCRA) 1956. Today, BSE is the world's number 1
exchange in terms of the number of listed companies and the world's 5th in handling of
transactions through its electronic trading system. The companies listed on BSE command a total
market capitalization of USD Trillion 1.06 as of July, 2009. The BSE Index, SENSEX, is India's
first and most popular stock market benchmark index. Sensex is tracked worldwide. It constitutes
30 stocks representing 12 major sectors. It is constructed on a 'free-float' methodology, and is
sensitive to market movements and market realities. Apart from the SENSEX, BSE offers 23
indices, including 13 sectoral indices.
We have taken monthly share price data for sixty months (March 2003- February 2008) of just
five companies which are included in BSE index. Though any number of stocks can be
considered, we have taken only five stocks to reduce the complexity. The Table 5.1 shows the
companies name along with their return in the form of trapezoidal uncertain numbers.

Table 5.1 Stocks and their returns
Company Return ( ir% )
Reliance Energy (RE) (-0.008, 0.031,0.067)
Larsen and Toubro (LT) (-0.003, 0.043, 0.087)
Tata Steel (TS) (0.009, 0.030, 0.052)
Bharat Heavy Electricals Limited (BH) (-0.002, 0.036, 0.083)
State Bank if India (SB) (-0.010, 0.033, 0.079)
With respect to the above data, we consider the following portfolio selection model:
1 1 2 2 3 3 4 4 5 5
1 1 2 2 3 3 4 4 5 5
1 1 2 2 3 3 4 4 5 5
1 2 3 4 5
1 2 3 4 5
Maximize PSR[r x r x r x r x r x ]
Maximize Skew[r x r x r x r x r x ]
subject to
( 5.1)
E[r x r x r x r x r x ] 0.04
x ,x ,x ,x ,x 0
x +x +x +x +x 1
+ + + +

+ + + +


+ + + + ≥
 ≥

=
% % % % %
% % % % %
% % % % %
We apply theorem 3.1 to convert model (5.1) into the deterministic model (3.7). To solve it, the
proposed MOGA is used. The cross over and mutation probabilities are chosen as 0.6 and 0.2
respectively. The number of iterations is 100. The solution is obtained as shown in Table 5.2.
Table 5.2 Portfolio
x1 x2 x3 x4 x5
0.3936170 0.00000 0.00000 0.6000000 0.00638296
Table 5.2 shows that the investor should invest 39.36%, 60% and 0.64% of the total money in the
1st
, 4th
and 5th
stocks. The portfolio is explained via the pie chart given in Figure 5.2.
Figure 5.2 Portfolio

6. CONCLUSIONS
In this paper, a new framework of fuzzy portfolio selection is introduced. Instead of following the
return-risk framework, this work concentrates on the risk-aversion nature of the investors and set
up a possibilistic Sharpe ratio –skewness portfolio selection problem. To do so, the possibilistic
Sharpe ratio is defined. As the Sharpe ratio prefers symmetric distribution, we consider skewness
to get rid of this drawback. The model is tested on a data set collected from BSE.
In near future, we will apply these portfolio selection models and solution method to other asset
allocation problems, combinational optimization models and multi-period problems to find
optimal investment strategy under complex market situations. Some other algorithms such as
ACO (ant colony optimization), PSO (particle swarm optimization), VEGA (vector evaluation
genetic algorithm), NEGA (Nondominated sorting genetic algorithm), NPGA (Niched Pareto
genetic algorithm) and PAES (Pareto archived evolution strategy) may be employed to solve the
problem, especially when the data set is significantly large.
REFERENCES
[1] H. Markowitz, Portfolio selection, Journal of Finance 7(1952) 77–91.
[2] T. Lai, Portfolio selection with skewness: a multiple – objective approach, Review of the Quantitative
Finance and Accounting 1 (1991) 293-305.
[3] H. Konno, H. Shirakawa, A mean- absolute deviation-skewness portfolio optimization model, Annals
of Operations Research 45 (1993) 205-220.
[4] H. Konno, K. Suzuki, A mean-variance-skewness optimization model, Journal of the Operations
Research Society of Japan 38 (1995) 137- 187.
[5] P. Chunhachinda, P. Dandapani, S. Hamid, A. J. Prakash, Portfolio selection and skewness: evidence
from international stock markets, Journal of Banking and Finance 21 (1997) 143–167.
[6] S. C. Liu, S. Y. Wang, W. H. Qiu, A mean- variance- skewness model for portfolio selection with
transaction costs, International Journal of System Science 34 (2003) 255- 262.
[7] A. J. Prakash, C. Chang, T. E. Pactwa, Selecting a portfolio with skewness: recent evidence from US,
European and Latin American equity markets, Journal of Banking and Finance 27 (2003) 1375- 1390.
[8] W. Briec, K. Kerstens, O. Jokung, Mean-variance- skewness portfolio performance gauging: a general
shortage function and dual approach, Management Science 53 (2007) 135- 149.
[9] L. Yu, S. Y. Wang, K. Lai, Neural network based mean- variance-skewness model for portfolio
selection, Computers and Operations Research 35 (2008) 34 – 46.
[10] R. Bhattacharyya, M. B. Kar, S. Kar, D. Dutta Majumder, Mean- entropy- skewness fuzzy portfolio
selection by credibility theory approach, In: S. Chaudhury et al. (Eds.): PReMI 2009. LNCS 5909, pp-
603–608, 2009.
[11] Bhattacharyya, R.; Kar, S. Possibilistic mean-variance-skewness portfolio selection models.
International Journal of Operations Research 8 (2011) 44 – 56.
[12] R. Bhattacharyya, S. Kar, D. Dutta Majumder, Fuzzy mean–variance–skewness portfolio selection
models by interval analysis, Computers and Mathematics with Applications 61 (2011) 126-137.
[13] A. Chatterjee, R. Bhattacharyya, S. Mukherjee, S. Kar, Optimization of mean-semivariance-skewness
portfolio selection model in fuzzy random environment, In: ICOMOS 2010, American Institute of
Physics conference proceedings, pp- 516-521.
[14] H. Markowitz, Portfolio selection: efficient diversification of investments, New York: Wiley, 1959.
[15] A. D. Roy, Safety first and the holding of assets, Econometrics 20 (1952), 431– 449.
[16] G. Castellacci, M. J. Siclari, The practice of delta–gamma VaR: implementing the quadratic portfolio
model, European Journal of Operational Research 150 (2003) 529–545.
[17] J. Philippe, Value at risk: the new benchmark for controlling market risk, Irwin Professional
Publishing, Chicago, 1996.
[18] G. C. Philippatos, C. J. Wilson, Entropy, market risk and selection of efficient portfolios, Applied
Economics 4 (1972) 209 – 220.

[19] G. C. Philippatos, N. Gressis, Conditions of equivalence among E–V, SSD, and E–H portfolio
selection criteria: The case for uniform, normal and lognormal distributions, Management Science 21
(1975) 617–625.
[20] D. N. Nawrocki, W. H. Harding, State-value weighted entropy as a measure of investment risk,
Applied Economics 18 (1986) 411–419.
[21] M. R. Simonelli, Indeterminacy in portfolio selection, European Journal of Operational Research 163
(2005) 170-176.
[22] X. Huang, Mean-entropy models for fuzzy portfolio selection, IEEE Transactions on Fuzzy Systems
16(2008) 1096–1101.
[23] Z. Qin, X. Li, X. Ji, Portfolio selection based on fuzzy cross-entropy, Journal of Computational and
Applied Mathematics, 228(2009) 188- 196.
[24] W. F. Sharpe, Mutual fund performance, Journal of Business, 39 (1966) 119 – 138.
[25] H. Konno, K. Suzuki, A mean-variance-skewness optimization model, Journal of the Operations
Research Society of Japan 38 (1995) 137- 187.
[26] T. Leon, V. Liern, E. Vercher, Viability of infeasible portfolio selection problems: a fuzzy approach,
European Journal of Operational Research 139 (2002) 178-189.
[27] E. Vercher, J. D. Bermudez, J. V. Segura, Fuzzy portfolio optimization under downside risk
measures, Fuzzy Sets and Systems 158(2007) 769 – 782.
[28] H. Tanaka, P. Guo, Portfolio selection based on upper and lower exponential possibility distributions,
European Journal of Operational Research 114(1999) 115 – 126.
[29] M. Inuiguchi, T. Tanino, Portfolio selection under independent possibilistic information, Fuzzy Sets
and Systems 115(2000) 83–92.
[30] K. K. Lai, S. Y. Wang, J. P. Xu, S. S. Zhu, Y. Fang, A class of linear interval programming problems
and its application to portfolio selection, IEEE Transactions on Fuzzy Systems 10(2002) 698–704.
[31] S. Y. Wang, S. S. Zhu, On fuzzy portfolio selection problems, Fuzzy Optimization and Decision
Making 1(2002) 361- 377.
[32] S. Giove, S. Funari, C. Nardelli, An interval portfolio selection problems based on regret function,
European Journal of Operational Research 170(2006) 253–264.
[33] M. Ida, Solutions for the portfolio selection problem with interval and fuzzy coefficients, Reliable
Computing 10(2004) 389–400.
[34] C. Carlsson, R. Fuller, P. Majlender, A possibilistic approach to selecting portfolios with highest
utility score, Fuzzy Sets and Systems 131(2002) 13–21.
[35] Y. Fang, K. K. Lai, S. Y. Wang, Portfolio rebalancing model with transaction costs based on fuzzy
decision theory, European Journal of Operational Research 175(2006) 879– 893.
[36] X. Wang, W. J. Xu, W. G. Zhang, M. L. Hu, Weighted possibilistic variance of fuzzy number and its
application in portfolio theory, Lecture Notes in Artificial Intelligence 3613(2005) 148–155.
[37] W. G. Zhang, Y. L. Wang, Portfolio selection: possibilistic mean-variance model and possibilistic
efficient frontier, Lecture Notes in Computer Science 3521(2005) 203–213.
[38] V. Lacagnina, A. Pecorella, A stochastic soft constraints fuzzy model for a portfolio selection
problem, Fuzzy Sets and Systems 157(2006) 1317–1327.
[39] C. Lin, B. Tan, P. J. Hsieh, Application of the fuzzy weighted average in strategic portfolio
management, Decision Sciences 36(2005) 489–511.
[40] X. Huang, Two new models for portfolio selection with stochastic returns taking fuzzy information,
European Journal of Operational Research 180(2007) 396–405.
[41] W. Fei, Optimal consumption and portfolio choice with ambiguity and anticipation, Information
Sciences 177(2006) 5178– 5190.
[42] W. G. Zhang, Y. L. Wang, Z. P. Chen, Z. K. Nie, Possibilistic mean-variance models and efficient
frontiers for portfolio selection problem, Information Sciences 177(2007) 2787-2801.
[43] J. Li, J. Xu, A class of possibilistic portfolio selection model with interval coefficients and its
application, Fuzzy Optimization and Decision Making 6(2007) 123–137.
[44] M. A. Parra, A. B. Terol, M. R. Urfa, A fuzzy goal programming approach to portfolio selection,
European Journal of Operational Research 133(2001) 287- 297.
[45] A. B. Terol, B. P. Gladish, M. A. Parra, M. R. Urfa, Fuzzy compromise programming for portfolio
selection, Applied Mathematics and Computation 173(2006) 251 – 264.
[46] X. Huang, Mean-semivariance models for fuzzy portfolio selection, Journal of Computational and
Applied Mathematics 217(2008), 1–8.
[47] X. Huang, Risk curve and fuzzy portfolio selection, Computers and Mathematics with Applications
55 (2008) 1102– 1112.

[48] L. A. Zadeh, Fuzzy sets as a basis for a theory of possibility, Fuzzy Sets and Systems 1 (1978) 3–28.
[49] D. Dubois, H. Prade, The mean value of a fuzzy number, Fuzzy Sets and Systems 24 (1987) 279–300.
[50] C. Carlsson, R. Fuller, On possibilistic mean value and variance of fuzzy numbers, Fuzzy Sets and
Systems 122 (2001) 315–326.
[51] R. Fuller, P. Majlender, On weighted possibilistic mean and variance of fuzzy numbers, Fuzzy Sets
and Systems 136 (2003) 363–374.
[52] B. Liu, Y. K. Liu, Expected value of fuzzy variable and fuzzy expected value models, IEEE
Transactions on Fuzzy Systems 10 (2002) 445-450.
[53] A. Saeidifar, E. Pasha, The possibilistic moments of fuzzy numbers and their applications, Journal of
Computation and Applied Mathematics 223(2009) 1028 – 1042.
[54] I. Georgescu, Possibilistic risk aversion, Fuzzy Sets and Systems 160 (2009) 2608 – 2619.
[55] R. Bhattacharyya, P. Kumar, S. Kar, Fuzzy R&D portfolio selection of interdependent projects,
Computers and Mathematics with Applications 62 (2011) 3857–3870.
[56] A. Roy, S. Kar, M. Maiti, A deteriorating multi-item inventory model with fuzzy costs and resources
based on two diﬀerent defuzziﬁcation techniques, Applied Mathematical Modelling 32 (2008) 208 –
223.
AUTHORS
Dr. Rupak Bhattacharyya is presently working as an Associate Professor in Mathematics
and HOD in the Department of Applied Science & Humanities of Global Institute of
Management & Technology, Krishnagar, West Bengal, India. Dr. Bhattacharyya received
his PhD degree from National Institute of Technology, Durgapur. The Operational
Research Society of India confers the Professor M N Gopalan Award on Dr.
Bhattacharyya for the best Doctoral Thesis in Operational Research of 2011. He has more
than eight years of teaching and research experience. He has more than 20 International
publications. Details on Dr. Bhattacharyya can be viewed from: https://sites.google.com/site/rupakmath.

POSSIBILISTIC SHARPE RATIO BASED NOVICE PORTFOLIO SELECTION MODELS

More Related Content

What's hot (18)

Similar to POSSIBILISTIC SHARPE RATIO BASED NOVICE PORTFOLIO SELECTION MODELS (20)

More from cscpconf (20)

Recently uploaded (20)

POSSIBILISTIC SHARPE RATIO BASED NOVICE PORTFOLIO SELECTION MODELS