||A Problem Sohing Agent (PSA) is either an active agent (e.g. a human) or a passive agent (e.g. a hardware or software system) having an ability to execute a finite set of tasks in an application domain. An Activity represents any abstract description of works in the application domain, and we refer to it as a problem that needs cooperation and interaction among the PSAs to complete it. A Tusk is an atomic activity that is executed by exactly one PSA. An Activity Munagement System (AMS) deals with the specification, management and execution of activities with the cooperation and interaction among various Problem Solving Agents (PSAs). Each PSA has its competence defined by a set of abilities it has to execute tasks. Each activity requires a certain competence from PSAs specified as a set of needs for executing all of its tasks. A token is an unique fixed-length character string that embodies a capability describing the ability of a PSA to execute certain functions/procedures/tasks. The needs of a task are also modeled by the token. An activity can be decomposed into multiple inter- dependent tasks to be coordinated and scheduled by events. It is possible that a batch of potential PSAs may be selected for a task. Therefore various policies are developed to select one candidate PSA from this batch. The PSAs are selected by applying match making policies using the tokens (which are managed by the Capability Database). Since PSAs can achieve similar capabilities in different ways, we develop a token equivalence theory and partial match making. This token equivalence theory is used to augment the match making process in selecting a PSA for a task. In this thesis, we present a framework for capability-based activity specification and decomposition of AMS. We also provide an intermediate language generator to generate a script to describe an activity for further activity execution and monitoring by the AMS. Lastly, we present the issues considered in the prototype implementation of activity decomposition , specification and match making components of an AMS and future work.