Automatic Conflict Detection on Contracts

Abstract

Abstract. Many software applications are based on collaborating, yet competing, agents or virtual organisations exchanging services. Contracts, expressing obligations, permissions and prohibitions of the different actors, can be used to protect the interests of the organisations engaged in such service exchange. However, the potentially dynamic composition of services with different contracts, and the combination of service contracts with local contracts can give rise to unexpected conflicts, exposing the need for automatic techniques for contract analysis. In this paper we look at automatic analysis techniques for contracts written in the contract language CL. We present a trace semantics of CL suitable for conflict analysis, and a decision procedure for detecting conflicts (together with its proof of soundness, completeness and termination). We also discuss its implementation and look into the applications of the contract analysis approach we present. These techniques are applied to a small case study of an airline check-in desk.

CLAN: A Tool for Contract Analysis and Conflict Discovery

“…Furthermore, other contract analysis techniques are planned to be added to the tool to enable analysis for overlapping, superfluous and unreachable sub-clauses of a contract. See [1] for related works.…”

“…We are currently working on how to optimise the analysis to avoid such exponential state-explosion. 1. The ground crew is obliged to open the check-in desk and request the passenger manifest two hours before the flight leaves.…”

“…The language has a trace semantics [2], which although useful for runtime monitoring of contracts, lacks the deontic information concerning the obligations, permissions and prohibitions of the involved parties in the contract, and thus it is not suitable for conflict analysis. We have recently developed conflict analysis techniques for CL [1], and in this paper, we present a tool implementing these techniques for the automatic analysis of contracts written in CL.…”

“…The above says that empty traces satisfy an obligation, while non-empty traces satisfy the contract if either (i) the obligation is satisfied, and the obligation is registered in the deontic trace; or (ii) the reparation (CTD) is satisfied by the remainder of the trace. See [1] for more details.…”

Evaluating how agent methodologies support the specification of the normative environment through the development process

“…It is necessary to ensure that each single normative environment has no conflicts, and also that the composition of all the normative environments is itself conflict-free. As is presented in [32], conflicts in norms arise for four different reasons: (1) the obligation and prohibition to perform the same action; (2) the permission and prohibition to perform the same action; (3) obligations of contradictory actions; (4) permissions and obligations of contradictory actions. Therefore, guidelines for verifying the coherence of the normative environment should be offered by the methodology and integrated into the development process.…”