| US 7,552,078 B2 | ||
| Enterprise portfolio analysis using finite state Markov decision process | ||
| David Flaxer, Dobbs Ferry, N.Y. (US); Paul Gregory Greenstein, Croton-on-Hudson, N.Y. (US); Robert Cornelius Hampshire, Lima, Ohio (US); Anil Nigam, Stamford, Conn. (US); and John George Vergo, Yorktown Heights, N.Y. (US) | ||
| Assigned to International Business Machines Corporation, Armonk, N.Y. (US) | ||
| Filed on Feb. 28, 2005, as Appl. No. 11/67,219. | ||
| Prior Publication US 2006/0195373 A1, Aug. 31, 2006 | ||
| Int. Cl. G06Q 40/00 (2006.01) | ||
| U.S. Cl. 705—36R [705/35; 705/37] | 3 Claims |

| 1. A computer implemented method for generating a transition plan for an enterprise portfolio of computer resources, the computer
performing the steps of:
storing in a computer memory an inventory data representing each of a plurality of computer resources available to an enterprise,
wherein said inventory data includes, for each of said computer resources, a baseline set of current computer resource capabilities,
wherein each of said computer resources has an associated state data representing its state within a life cycle state space
including a develop state, a steady state and a dead state;
providing a set of enterprise strategic goals for transitioning the enterprise portfolio of computer resources;
translating said enterprise strategic goals into a set of future computer resource requirements for at least one of said computer
resources;
forming an initial transition plan, comprising actions pertaining to at least one of said computer resources, for transitioning
from said current computer resource capabilities to said future computer resource requirements, based on said current computer
resource capabilities and said future computer resource requirements, wherein said actions include a retain action, a retire
action and a restructure action, and wherein said actions transition the associated state data of said computer resource among
states within said life cycle state space;
generating an optimized transition plan of actions pertaining to at least one of said computer resources for transitioning
from said current computer resource capabilities to said future computer resource requirements, said generating including
optimizing said initial transition plan; and
outputting the optimized transition plan,
wherein said translating said enterprise strategic goals into one or more future computer resource requirements is based on
the expression
{Sj(Xin,ain)|0≤j≤K}
where
n represents a specific time period among said N time periods,
Sj represents said future computer resource requirements,
Xin represents the state, within said life cycle state space, of an i-th resource in said enterprise portfolio of computer resources
at time n,
ain represents an action applied to said i-th resource at time n and,
K represents a total number of said computer resources, and
wherein said optimizing said initial transition plan includes selecting a sequence of actions based on maximizing a business
value for said enterprise portfolio of computer resources, wherein said maximizing bases the business value of each computer
resource on:
the state of said computer resource within said life cycle state space,
a set of actions for transitioning said computer resource among states within said life cycle state space, and
a reward for each of said actions, wherein said reward is a cost associated with carrying out said action;
wherein said optimizing initial transition plan includes a backward inductive maximization to maximize said business value
for said enterprise portfolio of computer resources based on the algorithm:
![]() ā=(ān)|0≤n≤N)
where,
ā represents a portfolio transition plan as a set of portfolio actions for the an entire time horizon, each portfolio action
being a set of actions, each action corresponding to at least one computer resource of said enterprise portfolio,
JāN represents a cumulative adjusted business value,
N represents a number of time periods considered for said transition plan,
XjN represents an iterative state of a life cycle of an i-th resource in said enterprise portfolio of computer resources,
Zn is a risk adjusted business value at each decision period, and
ān represents portfolio actions to be performed at the n-th time period.
|