Feel free to enquire about hardcopies and downloading problems regarding papers in this highly personal selection, annotated by Magnus Boman.
I start with stuff moved from my recent papers page because they stopped being recent:
My masters thesis Boman, M.: Formal representation of the reasoning of intelligent agents by means of epistemic and doxastic modal propositional logic (DSV Report WP153, Stockholm Univ, 1989) looked at Robert Moore's work, and combined my undergraduate studies in logic with my attempt to delve into the artificial intelligence literature. There were not many papers with "intelligent agent" in the title around this time, and I was thrilled to find other people interested in 1990, when I attended a highly inspiring conference in Keele, UK. There, I met several people that helped found the area of multi-agent systems; among these were Eric Werner, Cristiano Castelfranchi, and Julia Galliers. Julia bravely invited me to Cambridge, and Eric got me involved with MAAMAW, the first European agent event, which I still like a lot.
I worked in a project on federated databases, and co-wrote several papers with Paul Johannesson during this time, including Boman, M. and Johannesson, P.: Axiomatic epistemic systems for federated information systems (In Prakash, N., ed., Data Management, 195--213. McGraw-Hill, 1990) and Boman, M. and Johannesson, P.: Epistemic logic as a framework for federated information systems (In Deen, M., ed., Cooperating Knowledge Based Systems, 255--270. Springer-Verlag, 1990) . I used parts of these in my doctoral thesis A logical specification of federated information systems (DSV Report 93-019-DSV, Stockholm Univ, 1993, Abstract) . The book was on agents, basically, but this time political reasons nixed me to put "intelligent agent" in the title, leading to a small circle of readers (my guess is three!). Judging by correspondence and references, the number of readers have since increased, possibly because the thesis contains two innovative ideas. One is that 3 sometimes equals 4. The other is a form of propositional truth-table that I have not seen elsewhere. Towards the end of my thesis work, one of my supervisors Per-Erik Malmnäs told me I had to choose between provability logic and all my other interests. I took the latter route, which I only regret whenever I meet a brilliant logician interested in provability, usually Ian Hodkinson. If you are curious about provability logic, have a look at Boman, M.: A survey of provability logic (DSV Internal Working Note 46, 1992, PDF [102K]) .
I returned to Keele for the follow-up in 1994, where I published the explanation of the 3=4 mystery in Boman, M. and Ekenberg, L.: Eliminating paraconsistencies in 4-valued cooperative deductive multidatabase systems with classical negation (In Deen, M., ed., Proc Cooperating Knowledge Based Systems, 161--176. Keele Univ Press, 1994, Postscript [91K]) , completed with the help of Love Ekenberg, with whom I was to co-lead the DECIDE research group a few years later. Love and I wrote many papers together, sometimes with DECIDE colleague Mats Danielson, many of which are downloadable from the DECIDE WWW-pages.
Representative of our decision analysis approach to agent design are Ekenberg, L.; Danielson, M. and Boman, M.: From local assessments to global rationality (Intl Journal of Intelligent and Cooperative Information Systems 5(2&3): 315-331, 1996, Postscript [651K]) (which by the way is better than our ICMAS95 paper on the same topic) and Ekenberg, L.; Danielson, M. and Boman, M.: Imposing security constraints on agent-based decision support (Decision Support Systems Intl Journal 20(1): 3-15, 1997, Postscript [296K]) . Our best paper by far though, grew out of Love's stay at IIASA, which we both still cooperate with. In Ekenberg, L.; Boman, M. and Linnerooth-Bayer, J.: General Risk Constraints (Journal of Risk Research 4(1): 31-47, 2001, PDF [176K]) , we characterise the class of constraints that can and should augment the principle of maximising the expected utility when representing risk. We more or less finished the paper already in 1997, but it was published only recently.
Besides my work with Love and IIASA, I wrote papers on agents that deviated less from what was expected in that area. Many of these were written with the help of my students, including Harko Verhagen who has many of our papers available on his WWW-pages. I published on norms in multi-agent systems, inspired by discussions with Cristiano Castelfranchi, see e.g.: Boman, M.: Norms in artificial decision making (Artificial Intelligence and Law 7: 17-35, 1999, Postscript [263K]) . My use of the term norm, even when I tried to emphasise that I dealt with technical norms, irritated many people. The cause of concern was (and continue to be) that the term norm had other connotations. Although this has not stopped agent researchers from redefining rationality, trust, and many other concepts, I still wish I had called my concept something else. I learned the lesson, though, and dubbed my next term pronouncer, for stand-alone decision support algorithms. The first publication on pronouncers was Boman, M. and Verhagen, H.: Social intelligence as norm adaptation (Proc workshop on Socially Situated Intelligence, Dautenhahn and Edmonds eds., 5th International Conference on Simulation of Adaptive Behavior, 1998) . We then moved on to test pronouncers in the robotic soccer domain RoboCup.
I was team captain of a simulated team called UBU for several years. This task was towards the end taken over by my graduate student Johan Kummeneje, who wrote his licentiate thesis on RoboCup. The UBU team has its own pages with downloadable publications. I also started up a Swedish national team of physical robots that competes in the four-legged Sony robot league. The current team captain for Team Sweden is Alessandro Saffiotti, who maintains information pages on past, present, and future activities.
Together with colleagues in Ronneby, I also investigated the use for norms, pronouncers, and agents in general in the domain of intelligent buildings. My co-author on virtually all these papers are Paul Davidsson, who maintains a WWW-page of papers, many of which are downloadable. A paper which I think neatly describes the use of pronouncers is: Boman, M.; Davidsson, P. and Younes, H.: Artificial decision making under uncertainty in intelligent buildings (Proc UAI'99, Laskey & Prade eds., pp.65-70, Morgan Kaufmann, 1999, Postscript [1207K]).
After RoboCup, I could not stay away from competitions entirely, and I am since the spring of 2000 involved in the Trading Agent Competition. The connection between computer science and economics is to me very interesting and important, and while we have not yet published on our current TAC team, I have written several papers on agents and e-commerce recently, including some with professor Koen Bertels who in 2000 visited me for part of his sabbatical, e.g.:
Bertels, K. and Boman, M.: Agent-Based Social Simulation in Markets (Electronic Commerce Research 1(1-2): 149-158, 2001, Postscript [864K])
. I co-advice Stefan Johansson, with whom I have published on agent-based simulations relating to both economics and biology/physics; a recent paper is:
Boman, M.; Johansson, S. and Lybäck, D.: Parrondo strategies for artificial traders (Presented at IAT'01, October 2001)