This thesis concerns the integration of agent technology and mathematical optimization for improved decision support within the domain of analysis and planning of production and transportation. These two approaches have often been used separately in this domain but the research concerning how to combine them is very limited. The studied domain is considered to be complex due to the fact that many decision makers, which influence each other, often are involved in the decision making process. Moreover, problems in the domain are typically large and combinatorial, which makes them more difficult to solve. We argue that the integration of agent-based approaches and mathematical optimization has a high potential to improve analysis and planning of production and transportation. In order to support this hypothesis, we have developed and analyzed three different approaches to the integration of agent technology and mathematical optimization. First, we present a Multi-Agent-Based Simulation (MABS) model called TAPAS for simulation of decision-making and physical activities in supply chains. By using agent technology and optimization, we were able to simulate the decision-making of the involved actors as well as the interaction between them, which is difficult using traditional simulation techniques. In simulation experiments, TAPAS has been used to study the effects of different types of governmental taxes, and synchronization of timetables. Moreover, we provide an analysis of existing MABS applications with respect to a number of criteria. Also, we present a framework containing a number of abstract roles, responsibilities, and interactions, which can be used to simplify the process of developing MABS models. Second, we present an approach for efficient planning and execution of intermodal transports. The approach provides agent-based support for key tasks, such as, finding the optimal sequence of transport services (potentially provided by different transport operators) for a particular goods transport, and monitoring the execution of transports. We analyzed the requirements of such an approach and described a multi-agent system architecture meeting these requirements. Finally, an optimization model for a real world integrated production, inventory, and routing problem was developed. For solving and analyzing the problem, we developed an agent-based solution method based on the principles of Dantzig-Wolfe decomposition. The purpose was to improve resource utilization and to analyze the potential effects of introducing VMI (Vendor Managed Inventory). In a case study, we conducted simulation experiments, which indicated that an increased number of VMI customers may give a significant reduction of the total cost in the system.