Putting APL platforms to the test: agent similarity and execution performance
Tristan M. Behrens, Koen V. Hindriks, Jomi Fred Hübner, Mehdi M. Dastani · 2010
It is our goal to compare agent programs, implemented by means of dif-ferent agent programming languages, by firstly reasoning about their sim-ilarity and secondly by measuring the time it takes them to reach specific states. In this paper, we will establish a similarity notion based on compar-ing agent states and agent runs. An agent state is usually defined as a snap-shot of the agent’s internals during execution, an agent run is usually defined as a sequence of such agent states. Examining the different no-tions of agent states specific to different agent programming languages, we will define generic agent states/runs and establish a similarity notion on top of that. After defining a set of mappings, that map specific agent states/runs to generic ones in order to facilitate a common starting point,