New England Complex Systems Institute

Special joint meeting March 27, 1998

The New England Complex Systems Institute (NECSI)
The Center for Computational Science (CCS) of Boston University

Information Mechanics

Time: Friday March 27, 10am - 5pm

Location: Physics Research Building, Boston University 3 Cummington St. Room 593, Next to 590 Commonwealth Ave.


Michel Baranger (Physics Dept., MIT) : Following entropy change and chaos in the transition to equilibrium
Norm Margolus (AI Lab, MIT) : Information and entropy in reversible cellular automata
Jeff Yepez (Hanscom Air Force Laboratory): Lattice Gas Theory and Computation
Charles Bennett (IBM): Tutorial --- Algorithmic complexity and logical depth

This meeting will be informal, it is organized as the second meeting of the New England Complex Systems Institute working group on Information Mechanics. Faculty and students are welcome to participate.

Lunch: Given the opportunities for convenient and inexpensive lunches in the BU area, we will forgo ordering a lunch and take to the streets!

Participation: Please notify Rick Bagley of your intention to attend:, who will also field your questions.

The format of these working group meetings is responsive to the interests of the participants. The meetings may include:

- formal seminars
- descriptions of ongoing research
- speculative presentations
- tutorials on fundamentals or primers on advanced topics
- overviews
- selection and study of published research

Ongoing discussion is the mortar for these elements.

The immediate benefit is to the work and study of participants! But elements of the presentations and discussions will be included in a NECSI publication volume.Additional information about working groups at NECSI on other topics can be found at and general information about necsi can be found at (or contact, Yaneer Bar-Yam, president of NECSI at


By Subway: The CCS is right on the Green Line. You can actually walk to it from Kenmore Square, but if you can get on the Green Line that goes out of Kenmore Square along Commonwealth Avenue, then the very first stop after Kenmore Square, called Blandford, is even better. This is located at the intersection of Comm. Ave. and Blandford St. From there you can actually see the BU Physics Research Building ("PRB" in the map below) at the corner of Blandford and Cummington. Its address is 3 Cummington.

By car: If you are driving to Boston, it is convenient to park at one of the extremal points of the "T", such as at Alewife on Route 2, and continue on as outlined above. Contact Rick Bagley if you need help finding such spots.

For those who will drive in to the university, here's a road map:

                  Commonwealth Avenue
  ============================================= Kenmore
                            |                   Square
                       *PRB*|Blandford St.      /
     -----------------------|          ________/
        Cummington St.                /
                   ________/  Beacon St.

Hopefully, you can find these streets on a real map of Boston. (No question about Commonwealth and Beacon; you'll need a detailed map to see Blandford and Cummington, however.) Good luck finding a meter; once you do, don't let it run out of time, as the meter maids here operate with deadly efficiency.

If you enter the building from Cummington St., there will be elevators right inside the entrance. Take one to the fifth floor. Take a left out of the elevator, pass through double doors, and then take another left. Pass by the restrooms and the computer rooms, and PRB 593 is the last room on the left.


Jeffrey Yepez, Air Force Research Laboratory

Presented in this talk are a selection of results from our lattice gas research that has been ongoing for five years at the Air Force Research Laboratory located at Hanscom AFB in Massachusetts. A lattice gas is a system of indistinguisable particles moving and colliding on a discrete spacetime lattice. Computational simulations carried out at the Air Force Research Laboratory on the CAM-8, a lattice gas machine, verify theoretical results showing the fluid-like behavior of lattice gases at macroscopic scales.

A mesoscopic lattice Boltzmann method for hydrothermal fluids and multiphase fluids are briefly discribed. However, our focus has been on microscopic lattice gases, viscous Navier-Stokes fluids and complex fluids, such as liquid-gas systems and microemulsions. We have explored two generalizations of the classical lattice gas: (1) an integer lattice gas; and (2) a quantum lattice gas. Classical lattices gases use a single classical bit to encode the presence of a particle, whereas integer lattice gases and quantum lattice gases use an integer and a qubit, respectively, to encode particle occupancies.

For the integer lattice gas system, we have developed a statistical mechanical treatment based on a lattice-gas Hamiltonian that allows us to compute the partition function in the grand canonical ensemble. Using a diagrammatic method, a Mayer's cluster expansion is performed to perturbatively calculate the equation of state of a multiphase lattice gas fluid in equilibrium.

In a quantum lattice gas system, at any particular time, many different particle configurations simultaneously exist on the lattice by quantum mechanical superposition of states. We have found quantum lattice gas algorithms for several physical systems: (1) the nonrelativistic many-body Schrodinger equation; (2) the Landau superfluid equations; and (3) the Navier-Stokes equation. Quantum lattice gases are clearest example illustrating the vast computational power of a quantum computer that grows exponentially in the number of qubits.