شبیه سازی برنامه ریزی پویا پرتفولیو پروژه

A main challenge in project management is to provide methodologies that facilitate coordination among the projects in a portfolio or in a firm. Each incoming project in an existing portfolio affects its schedule, the resources availability and the planned performance. There are no analytical solutions for the problem of dynamic scheduling of resources for multiple projects in real time. Mathematical approaches, like integer programming or network based techniques, cannot describe the complexity of real problems (multi-projects environments have many interrelated elements), and have difficulties to adapt the analysis to dynamic changes. We propose a multi-agent system, where projects negotiate the procurement of resources through an auction mechanism all over the portfolio life. Both, projects and resources are modelled as agents. Projects demand resources for fulfilling their scheduled planned work, whereas resources offer their capabilities and workforce. This framework allows project portfolio management and the assessment in the decision of acceptance/rejection new projects.

In the XXI century firms are adopting projects, programs and portfolios to meet their strategic, operational and tactical goals. Providing methodologies suitable for dealing with simultaneous management of multiple projects –differing in terms of size, required skills and urgency – remains as a main challenge for project management [25]. Firms create portfolios looking for benefits (synergies) from global coordination among projects, but some issues are not yet solved, as the dynamic multi-project scheduling. The decision to accept a new project mainly depends on its economic feasibility, its strategic importance and its scheduling feasibility. But incoming projects affect and are affected by the rest of the projects in the portfolio. Once a new project is accepted and incorporated to an existing portfolio, we have to update the projects task schedules and the expected return to be ready for new project acquisitions. Previous decisions have high impact on the office’s profit. In order to achieve strategic goals it is important to give priority to projects, and to allocate activities to the most efficient workers at the appropriate time. Because of this, before executing projects it is advisable to make a schedule that optimizes the allocation of resources. Classical methods, based on mathematical programming, can help to make those decisions when problem complexity is low and the system organization is rather static. But these techniques are not flexible or robust enough, and have difficulties to consider many real factors. In addition, real environments undergo frequent changes (new resources, new technologies) that force to modify the scheduling system. These issues have motivated, in last years, the development of new proposals for improving scheduling and control in multi-project environments. The paradigm of multi-agent systems (MAS) can help to find solutions, especially in cases where some social behaviour emerges. This paper shows an agent based approach for online dynamic scheduling and control in multi-project environments that takes advantage of the ability of agents to negotiate and adapt to changing conditions. The MAS has basically two types of agents: projects managers and resources managers. Projects have scheduled work to be done by different resources. Resources are endowed with some capabilities (knowledge, work force, etc.) that are needed to do the work. Projects demand resources over time and resources offer their capabilities and time availability. There is an auction process, and the price of resource-time slots emerges endogenously as a result of supply and demand. The rest of the paper is organized as follows. In the next section we discuss the problem of multi-project scheduling and the applicability of agent based simulation. Section 3 shows the MAS we have developed to provide real time evaluation of projects and the portfolio scheduling problem. Agents participate in a ‘virtual market’ that allows an efficient allocation of resources for project tasks which is explained in Section 4. The simulation of different scenarios and results are described in Section 5. Finally, we summarize the main conclusions in Section 6.

Although project management literature has been mainly concerned with managing individual projects, in practice firms usually work in dynamic and complex multi-project environments. To deal with these complex environments, we use multi-agent systems. They allow capturing real complexity, and managing the dynamical issues of the environment. In our particular case of multi-project systems, we have defined three kinds of artificial agents: resource manager agents, project manager agents and the MAC agent (auctioneer). Such design enables us to distribute the management system into elemental components directly identifiable, so it is easier to establish scheduling and control distributed mechanisms based on market procedures. The system allocates resources to projects dynamically, and it decides about project acceptance/rejection taking into account its impact on the existing portfolio in terms of value, profitability, schedule and operational information. We also show how it is possible to discover which resources are the most valuable for the organization. Results show that the auction-based allocation mechanism improves schedules and resource flexibility to achieve more efficient performance. This approach contributes to fill the gap between the literature in portfolio project management – usually focussed on corporate strategy and finance – with the work in multi-project management – mainly concerned with operational issues, scheduling and resource allocation. This approach can be extended by including some issues observed in real multi-project environments and not considered in this preliminary model. Thus, the possibility of subcontracting a task could be considered, or the possibility to define external restriction on tasks, etc. Muti-agent based modelling makes easy to implement these issues. In order to prove the system capacity to manage flexibility, an interesting study could be to test the system in a real flexible scenario. In our current approach, iterative price adjustment is based on resources capacity conflicts. There are other methods, as we can see in [21], in which price-adjustment is based on task precedence conflicts. In future works, it should be interesting to implement both methodologies analysing their similarities and differences, and even designing a new mechanism that includes features of both approaches.