![]() |
The mobile WAVE model described has a rather long history.1. The first mobile programs (not called "agents" at that time) were introduced and implemented in heterogeneous computer networks in 1972-74 [4], where special interpreted control programs were roaming between computers while activating distributed procedures and dragging exchange data with them.
2. In 1976-85, these ideas were used in a design of a recursive 256-processor distributed computer with interpreted control programs, which was serially produced by industry and used for modelling in aerodynamics.
3. In 1983-84, mobile code was used for parallel navigation of semantic networks and an interpreter of a simple WAVE-like language was implemented in a string-processing language. The navigation was based on a hop (or go) operation changing program's position in a network, with the program tail applied in a new current node (or nodes, if replicated in a broadcast to many neighbours).
4. In 1985-87, an extended WAVE [5] was compiled into LISP, and a real semantic network knowledge base for an international artificial intelligence laboratory was implemented, reflecting staff, topics, geography, time, etc., and heterogeneous links between them.
5. In 1989-92, WAVE was interpreted in Lisp and later used in an intelligent network management project financed by Siemens/Nixdorf.
6. A European Patent [6] had been filed in 1990 and issued in 1993 on a highly parallel computer architecture processing distributed semantic nets by mobile agents in WAVE.
7. In 1992-1994, WAVE interpreter was implemented in C under Unix in Internet, being currently a public domain.
8. In 1991-93, a special interest group called Kyma (WAVE, in Greek) existed in Silicon Valley (nearby San Jose) developing these ideas, with a small subset of WAVE [5] implemented by Matt Young from Lockheed.
9. Currently WAVE language [7] is being specialized, in cooperation for the design of a new type of dynamic control of advanced mobile cellular networks. It is also being used for simulation of distributed battlefields [2]. High-performance intelligent WAVE chip is being designed from which any parallel and distributed control structures may be composed, driven by cooperative mobile agent societies. Another version of such a chip is envisaged for an intelligent distributed and dynamic 3D graphics.
References
[1] J. Darling, P. Sapaty, P. Borst, "Modeling virtual worlds using the WAVE technology", Video presentation at the 14th DIS Workshop, Orlando, FL, March 1996 (http://www.ee.surrey.ac.uk/Research/DKP/).
[2] P. S. Sapaty, M. J. Corbin, S. Seidensticker, "Mobile intelligence in distributed simulations", Proc. 14th DIS Workshop, IST UCF, Orlando, FL, March 1996.
[3] Sapaty, P. S., and Borst, P. M. "WAVE: Mobile Intelligence in open networks", Proc. etaCOM'96, Portland, Oregon, May 1996.
[4] Sapaty, P. S., On possibilities of a direct intercomputer dialogue in ANALYTIC and FORTRAN languages, Publ. No. 74-29, Institute of Cybernetics Press, Kiev, 1974 (in Russian).
[5] "---", "WAVE-1: A new ideology of parallel processing on graphs and networks", Future Generations Computer Systems, Vol. 4, North-Holland, 1988.
[6] "---", "A distributed processing system", European Patent No. 0389655, Publ. 10.11.93, European Patent Office.
[7] "---", "Mobile WAVE technology for distributed knowledge processing in open networks", Proc. CIKM'95, Baltimore, MD, Dec. 1995 (http://www.cs.umbc.edu/~cikm/1995/npiv/sapaty/paper.ps).
[8] Vuong, S., Ivanov, I. "Mobile intelligent agent systems: WAVE vs. JAVA", Proc. etaCOM'96, Portland, Oregon, May 1996.
[9] S. Vuong, and L. Mathy, "Simulating the mobile-IP protocol using Wave", Proc. etaCOM'96, Portland, Oregon, May 1996.