Mario Brčić

Proactive Reactive Scheduling in Resource Constrained Projects with Flexibility and Quality Robustness Requirements

This paper presents a new approach to proactive reactive scheduling of stochastic resource-constrained project scheduling problems with known probability distributions of activity durations. To facilitate the search for cost-flexible proactive schedules that are adjustable and incur lower expected cost of future rescheduling, a new family of cost-based flexibility measures is introduced. Under these measures, cost is incurred on each rescheduling while taking into account the temporal distance of changes in the baseline schedule. We propose a new model that describes the integrated approach using the proposed cost-based flexibility measures where, in each stage, reactive scheduling can adjust the baseline schedule to accommodate flexibility and quality requirements. The model is based on bounded stochastic shortest path with finite state and action spaces. The commonly used schedule stability measure is put in the context of proposed family of flexibility measures and contrasted to them in the terms of project execution system properties.

Simulation library for Resource Constrained Project Scheduling with uncertain activity durations

Stochastic Resource Constrained Project Scheduling (SRCPS) is among the hardest combinatorial problems. Exact calculations of interesting measures, such as expected project duration and the probability of satisfying the deadline, using known probabilities are in #P even for relaxed instances of the problem where resource constraints are ignored. The most common approach is to use substantial simulation to evaluate candidate solutions. All of the work so far uses ad-hoc developed simulation environments with prevalent use of a priori generated activity duration scenarios. This paper describes the discrete-time simulation library aimed to support the creation of simulation-based algorithms for solving SRCPS problems with known probability distributions of activity durations. The library is designed, in the first instance, with shared memory parallelization of simulation, using OpenMP. Runtime parallelized generation of random activity duration scenarios is supported and we deal with “inconveniences” that are otherwise elegantly avoided using a priori generated activity duration scenarios. However, in some approaches that is not feasible and runtime scenario generation has to be used. We propose modular organization of simulators that enables better reuse of basic intrinsic project scheduling functionality.

Cloud-distributed computational experiments for combinatorial optimization

The development of optimization algorithms for combinatorial problems is a complicated process, both guided and validated by the computational experiments over the different scenarios. Since the number of experiments can be very large and each experiment can take substantial execution time, distributing the load over the cloud speeds up the whole process significantly. In this paper we present the system used for experimental validation and comparison of stochastic combinatorial optimization algorithms, applied in the specific case of project scheduling problems.

Planning horizons based proactive rescheduling for stochastic resource-constrained project scheduling problems

Abstract Parties that collaborate on projects need to synchronize their efforts. For this reason they seek a decreased rescheduling variability of the time arrangements. Proactive–reactive scheduling is important in such situations. It predominantly achieves synchronization through a shared baseline schedule and deviation penalties. As the latter currently introduce an unrealistically high level of inflexibility, the solution methods never proactively update the baseline schedule. We propose threshold-based cost functions for the deviation penalties to enable a more realistic modeling of aspects of project collaboration. These functions introduce a greater degree of flexibility through the notion of planning horizons for the activities. This results in the possibility of profitable proactive changes to the baseline schedule. We present two metaheuristic approaches for the case of stochastic durations: rollout-based and iterative policy search. Both these approaches use such opportunities to achieve substantial cost–performance improvements in comparison to the best existing method. This enhancement comes at the price of an increased computational burden and the greater complexity of the solution space.

Computational Aspects of Efficient Estimation of Harmonically Related Sine-Waves

This paper proposes an actual implementation of a well-known method [1] for spectral analysis of signals composed of harmonically related sine waves. The method itself requires computations which carried out directly according to the theoretical formulas do not yield computationally efficient implementation. Thus, utilizing matrix factorizations and mathematical “shortcuts”, several algorithms have been developed, which perform computations efficiently and make the method suitable for large-scale applications. Implementation details are clearly explained both theoretically and from computational point of view, and the achieved improvements have been proven by extensive simulations. Particular calculations applied will be equally efficient in all similar problems, which renders the proposed routines into widely useful building blocks.