SHAREX: A MULTIPERIOD PORTFOLIO MANAGEMENT MODEL

1
SHAREX A MULTIPERIOD PORTFOLIO MANAGEMENT MODEL
2
Key Features

• Integrated system of
• stock price forecasting
• portfolio optimization
• inventory management
• facilities for incorporating alternative
  techniques

3
Key Features

• the necessary financial relations included
• liquidity and debt, inventory, risk control
• transactions in discrete batch sizes
• fixed and variable transactions costs
• free specification of planning horizon
• forecasting and optimization combined
• extensive simulations for strategy specification
• real time management
• guaranteed feasibility

4
Large Scale Portfolio Management
5
Immediate research Topics

• parametric search under different economic
  conditions
• mixture density forecast models for skewed
  markets
• multicomputer implementation of SHAREX
• connections to efficient MINLP-solvers
• Utilizing VMA and IMA (volume/price index moving
  averages) in turning point detection

6
Background Research
Östermark, R. (1990) Portfolio Efficiency of
Capital Asset Pricing Models. Empirical Evidence
on Thin Stock Markets. Åbo Akademi University,
ISBN 951-649-703-9. Östermark, R. (1991) Vector
forecasting and dynamic portfolio selection.
European Journal of Operational Research 55,
46-56. Östermark, R Aaltonen J (1992)
Recursive Portfolio Management Large-Scale
Evidence from Two Scandinavian Stock Markets.
Computer Science in Economics and Management 5,
81-103. Östermark, R (2000a) A Hybrid Genetic
Fuzzy Neural Network Algorithm Designed for
Classification Problems Involving Several Groups.
Fuzzy Sets and Systems 1142, pp.
311-324. Östermark, R. (2000b) A Flexible Genetic
Hybrid Algorithm for Nonlinear Mixed-integer
Programming Problems. Accepted in
EvolutionaryOptimization.

7
Research (cont)
Östermark, R., Westerlund, T. Skrifvars, H.
(2000) A Nonlinear Mixed-Integer Multiperiod
Firm Model. International Journal of Production
Economics 67, p. 188-199. Östermark, R. (2001)
Genetic modelling of multivariate
EGARCHX-processes. Evidence on the international
asset return signal response mechanism.
Forthcoming in Computational Statistics Data
Analysis 38/1, 2001, pp. 1-124. Östermark, R.
(2002a) Automatic detection of parsimony in
heteroskedastic time series processes. Empirical
tests on global asset returns with parallel
geno-mathematical programming. Soft Computing
6/1, pp. 45-63. Östermark, R. (2002b) A
Multipurpose Parallel Genetic Hybrid Algorithm
for Nonlinear Non-convex Programming Problems.
Forthcoming in The European Journal of
Operational Research.