

The programs described in this site, must be installed as addins
for Microsoft Excel. This article explains how to install addins
and what to do in case of trouble. We also show how to get access
to the source code. 


This addin places the item Add ORMM on the OR_MM menu. Selecting
this item presents a dialog box that allows easy installation
and removal of the other addins in the collection. A second dialog
easily loads demonstration workbooks. 


The
Mathematical Programming Addin constructs optimization models
of several kinds:
The models can be solved using the Solver
Addin or one of the solution addins provided in this collection.
The Solver Addin comes with Excel, and it can solve linear
programming, integer programming and nonlinear programming
models. Additional addins are available in this package to
solve linear programming, integer programming, network programming
and transportation models.



This addin is an alternative to the Mathematical
Programming Addin for linear and integer programming models.
It's chief contribution is alternative format for models. The
formats are: Tableau, dual tableau, column list and row list.
The latter two store constraint coefficients in lists rather
than matrix. 


Stochastic programming
explicitly recognizes uncertainty by using random variables
for some aspects of the problem. With probability distributions
assigned to the random variables, an expression can be
written for the expected value of the objective to be optimized.
The expected value is to
be maximized or minimized.
For stochastic programming, some variables are to be set
by a decision maker, these are the decision variables,
while some model parameters are determined by chance, and
these are the random variables. To model a stochastic
programming problem, one must answer: When must the decision
maker make decisions relative to the time when the
random variables are realized? The several different answers
to this question lead to different computational methods.



This addin provides an algorithm that solves Linear Programming
or Integer Programming problems. It can be used instead of the
Excel solver for linear models created by the Mathematical Programming
addin. When the LP/IP Solver addin is installed a new item now
appears in the OR_MM menu, LP_Solver. Clicking on this item presents
a form allowing the selection three options: show a sensitivity
analysis, show detailed information about the steps primal simplex
procedure, and start the solution using the current solution value. 


A network solution algorithm is provided by this addin. The Excel
Solver actually solves network problems by solving the underlying
linear programming problem. Network algorithms are generally faster
than linear programming algorithms for solving problems that can
be modeled entirely as networks. The addin places a Network Solver
item on the OR_MM menu. In addition to pure network models, the
addin can solve generalized networks when arc gains are different
than 1. It can also solve models that require flows to be integer.
Clicking this item presents a dialog with which a number of solution
algorithm options are controlled. 


The Dynamic Programming Collection is a series of addins associated
with processes that involve states, actions and events.
Many situations can be described by a collection of mutually
exclusive states that are visited sequentially. From
each state the decision maker must choose an action.
Given the state and action the next state is determined by an event.
When the situation has only of states and events, the model
is a Markov
Chain. When the situation
has only of states and actions,
the model is a Deterministic Dynamic Program.
When the situation has states, actions and events, the model
is a Stochastic Dynamic Program, or Markov Decision
Process. The collection models and solves all of these
problems. 


The DP Examples section holds example problems, some
from the literature, that illustrate the capabilities of the Dynamic
Programming Collection. 


The DP Data addin provides the data structure for a
selected set of problems used to illustrate the remaining addins.
This addin has some interesting
problem classes of operations research and can
be revised to include new classes.
The DP Data addin will call the DP
Models addin and fill the forms created by
that addin. 


The DP Models addin constructs a form that
describes the states, actions and events characterizing a given
problem. It is an algebraic model generator similar to GAMS used
for Mathematical Programming models. The form constructed by
the DP Models addin holds the definitions of the states,
actions and events for the problem, formulas for computing the
objective function, and formulas for computing the transitions
from one state to another. The forms are filled by the DP
Data addin for certain problem classes. The DP Models addin
constructs lists of states, actions, events, decisions and transitions
that are used by the DP Solver addin. The DP Models addin
can also be used directly for problems not modeled by the DP
Data addin. 


The DP Solver addin creates a form
holding lists of states, actions, events, decisions and
transitions. The addin uses these lists with iterative
algorithms to find optimum actions for the states. The
addin handles, deterministic DP models, stochastic DP
models, and discrete time markov chains (DTMC) models.
Several solution strategies are provided.



Many Excel worksheet models depend on a few design variables.
Through formulas, the worksheet contents vary with the values
of the design variables and the modeler uses the worksheet as
a "what if" tool. It is often the goal to find the best
values. This addin provides algorithms that search for the best
among a perhaps large set of discrete alternatives. In addition
to standalone forms that can be used for general optimization
on a worksheet, the addin provides solutions to combinatorial
problems encountered in operations research studies. The problems
include mixed integer programming, the traveling salesman problem,
the assignment problem, spanning tree problems, optimum path problems
in networks and flowtree problems. 


This addin creates combinatorial models that use the search
methods of the Optimize addin. The addin provides a model
for the Quadratic Assignment problem, the Minimal Spanning
Tree problem and the Shortest Path Tree problem.
Other models may be added in the future. The Optimize
addin must be installed for the Combinatorics addin
to work. 


This program models and solves the vehicle routing problem
for several vehicles visiting several delivery sites. This
uses the methods of the Combinatorics addin, but the Routing
addin is a standalone program. A full description of the Routing
addin is in the OM/IE section
of this site. 


This addin performs operations on a function of continuous variables.
The function may depend on many variables and be constructed
of arbitrary combinations of Excel functions. The function
will usually be continuous and differentiable. The addin
uses numerical methods to compute the gradient and Hessian
matrices (matrices of first and second partial derivatives).
It also computes integrals and moments. The addin uses a
gradient search method to find values of the variables that
maximize or minimize the function. 


The Random
Variables Addin performs computations associated
with probability distributions. Random variables with any of 16
different named distributions may be defined. Functions compute
probabilities of events, inverse probabilities and moments. Up
to three distributions may be plotted. Complex probability problems
may be solved through MonteCarlo simulation. 


The Queuing
Addin performs calculations associated with Poisson queuing models,
Non Poisson models and networks of queues. It also performs simulations
of multichannel queuing models. 


This collection has been replaced by
the more general dynamic programming collection. 


This addin performs computations for discrete time or continuous
time Markov stochastic processes, DTMC or CTMC respectively.
The data defining a DTMC model are the probabilities
in the transition matrix, and the data defining a CTMC
model are the activity rates in the rate matrix. The addin
performs the following analyses: steady state probabilities,
nstep probabilities, first passage probabilities and expected
values, transient probabilities, simulation and absorbing state
probabilities. Economic data allows the performance of economic
analyses. 


A Markov Decision Process (MDP) adds decisions to
Markov analysis. Here states
may have several actions available to the decision maker. Actions
modify transition probabilities and have different costs or
rewards. The goal is to find an optimum policy that minimizes
expected discounted costs or maximizes discounted rewards.
This addin has been replaced by the more general DP Solver addin.



Consider a situation in which a series of decisions are to made
sequentially. The problem is complicated however in that the results
of some of the decisions are not deterministic, rather they are
affected by risk. The goal is to make a series of decisions in
order to maximize the expected return. Models and solution methods
for this type of problem are the subjects of decision analysis.
The addin provides extensive programs to build, solve and display
decision trees. 


This addin creates multiline simulations useful for analyzing
a variety of systems that don't fit the model types handled by
the other addins. The addin builds and maintains worksheets
on which simulations are easily built. Packaged models for time
series simulation and inventory simulation are included. 


This workbook tracks and forecasts hurricanes in the Gulf
of Mexico region. The download is an Excel workbook rather than
an addin, and it contains both sample data and the macros necessary
to add new hurricanes, add position data as provided by the National
Hurricane Center, make forecasts of future movement, plot data
and forecasts on a map and construct an error analysis after the
storm is over. 


To
train their production managers in the strategy developed,
the Proctor and Gamble company developed a simulation model
to illustrate the effects of the varying demand and how the
P&G strategy worked. At that time, the simulation was done
manually. The manual simulation was computerized
in the 1970’s in interactive BASIC and with the advent
of microcomputers became a viable tool in teaching Production
Scheduling and Inventory Control. More recently, it has been
converted by Paul Jensen to Visual Basic for Applications and
implemented for Microsoft Excel. We call the simulation the
P&G Game. 


Excel addins for several topics related to Operations Management
and Industrial Engineering are described in the OM/IE section
of this site.



