Probability Management Compatible Vendors

A Stochastic Information Packet or SIP is an array of simulated or historical realizations of an uncertain quantity. It provides a representation of a probability distribution as data with provenance, and is additive and auditable. The following software packages have been designed for compatibility with SIPs, and all are more powerful than native Excel for creating SIP libraries. We appreciate the ongoing support of the associated firms.

Below, we describe standards that will facilitate the use of libraries generated by such programs as inputs to SIP math dashboards in native Excel.

Analytica, from Lumina Systems

Powerful multi-dimensional modeling and simulation language. Calculates with the equivalent of SIP math internally and supports many distributions for performing Monte Carlo. Can import and export SIPs from and to Excel.

Crystal Ball, from Oracle Corp

Powerful traditional Monte Carlo Simulation for Excel, with a broad user base. Supports many distributions. Generates and reads SIPs.

Risk Solver Platform, Frontline Systems

Capable of interactive simulation and stochastic optimization on industrial sized problems, using both SIPs, and internally generated random numbers from many distributions.

XLSim, from Vector Economics

Free Excel add-in performs traditional Monte Carlo simulation with common distributions. Features include: SIP processing and DIST™ Distribution Strings, which compress SIP trials into a single cell for improved stochastic library management.

---

Stochastic Library Standard for Excel 10/30/2012

We have created a simple standard described here, and supported by the SIPmath Modeler Tools.



Stochastic libraries are stored in Excel worksheets containing SIPs and SLURPs (coherent sets of SIPs). The standard has three parts.

1. All the SIPs must have the same number of trials, and there must be a named cell, PM_Trials, containing the number of trials. When performing SIP math with such a library this number is used to dimension the data table for the simulation. That is, a column of integers from 1 through PM_Trials must be created somewhere in the workbook, with another cell specified to be the Column Input cell of the Data Table. The SIPmath Modeler Tool automates this by creating a new sheet in the model workbook, and gives the Column Input cell the name PM_Trials.
   
   
2. The SIPs can be in rows or in columns, but each one must be defined as a named range.
   
   
3. Optionally, there can be metadata Index Values appended to a SIP's trial values. These are useful for observing the results of your models under specified scenarios, such as average values of the inputs. If used, they are identified with the range named PM_IV containing the names of the metadata elements, and another named PM_IV_Index with the displacements to the metadata. These displacements must be common to all the SIPs in the library. For example, assuming a PM_Trials value of 1000 the two ranges might appear as below:
   
       
   
   Now if PM_Index is set to 1002, the Median of the SIPs will be displayed in all INDEX formulas. The values of the metadata must be specified when creating the library.