*** OR Society Local Search Study Group Talk ***

[Verity-Harrison Andrew <AVHARRISON@mail.dstl.gov.uk>]

*********

*** OPERATIONAL RESEARCH SOCIETY ***
*** LOCAL SEARCH STUDY GROUP TALK ***

Speaker: Dietmar Maringer, University of Essex

Title: Heuristic Optimisation for Portfolios with Transaction Costs

Time:	18:00-20:00, Thursday, 01/12/2005

Place:	Room 2005, Cass Business School, City University, 
	105 Bunhill Row, London EC1Y 8YZ

	http://www.cass.city.ac.uk/about/location/index.html

Abstract:

In portfolio theory, markets are mostly assumed to be perfect and 
frictionless. This allows focusing on the basic underlying mechanics and 
keeps the optimisation models simple and approachable. In real life, 
however, investors face all sorts of additional constraints, including 
integer constraints and transaction costs. Including these constraints would

bring the portfolio management models closer to real life problems -- yet 
exceeds the limits of many a traditional optimisation approach. To solve 
this dilemma between solvability and getting as close to real life problems 
as possible, an alternative type of optimisation techniques is needed, e.g.,

heuristic optimisation techniques.

A crucial feature of heuristic optimisation methods is that they do not (or 
not entirely) rely on deterministic procedures, but also incorporate 
stochastic elements. Typically, a heuristic optimisation techniques start 
off with a random solution and then iteratively suggest random changes which

are accepted or rejected according to some rule, which might be 
deterministic or, again, stochastic. The main advantage of this methods is 
that they are very flexible and therefore can be applied to all sorts of 
problems.

In this talk, I will introduce two of the simplest, but nonetheless highly 
effective heuristic optimisation techniques, namely Simulated Annealing and 
Threshold Accepting. These methods will then be compared to a more complex 
approach, Genetic Algorithms, and then be applied to the portfolio 
optimisation problem under transaction costs and integer constraints.

Our results imply that the heuristic optimisation is well apt to solve 
demanding problems in financial management -- and that the type transaction 
costs has a substantial effect on the portfolio structure.


Andy Verity-Harrison
 
Information Management Department
[dstl]
Rm D604
Malvern Technology Centre
 
Tel.   +44 (0)1684 771395         
Fax.   +44 (0)1684 771437
email: avharrison@dstl.gov.uk
 
****************************************************************************
************************************
Personal e-mail. This e-mail is personal. It is not authorised or sent on
behalf of the MOD. This e-mail is the personal responsibility of the sender.

Recipients should note that all e-mail traffic on MOD systems is subject to
monitoring and auditing

The remarks in this e-mail are personal and do not represent MOD or
Government policy.
****************************************************************************
************************************






"This e-mail is intended for the recipient only.  If you are not the
intended recipient you must not use, disclose, distribute, copy, print,
or rely upon this e-mail. If an addressing or transmission error has
misdirected this e-mail, please notify the author by replying to this e-mail."

"Recipients should note that all e-mail traffic on MOD systems is
subject to monitoring and auditing."