Meta-heuristics for wounded evacuation in disaster response

Supply Chain Management (SCM) is an essential element to operational efficiency, it requires the commitment of supply chain partners to work closely in coordinating order generation, order taking, and order fulfillment, Christopher et al. [1] Today, natural crises appear with increasing frequency, which causes floods, earthquakes, considerable fires, tsunamis, etc. In addition, artificial crises take their place, due to several regional and international factors, for example, the nuclear disaster of Chernobyl in 1986, the leaks in the Japanese nuclear reactors in Fukushima in 2011, the terrorist attacks (Paris-France, Tripoli -Lebanon, CairoEgypt) Etc. Several research projects have been conducted to develop decision support tools in postdisaster situations in urban areas and to provide accurate real-time information to the management center and rescue teams. All data collected, everywhere, in a disaster area, forms a database that can be relied on upon to make effective and, at the same time, fast rescue decisions. From some data, the human operator can easily make effective and accurate decisions. But with a very large amount of data (this is the case in urgent situations), we need to have a computer system capable to automatically process this data and efficiently sorts out important information that help the human agent to understand the situation and to quickly make the best relief decisions. Hence, a prompt disaster response, in addition to effective measures such as informed decision making, organized evacuation plan, right hospital selection, proper rescue vehicles, efficient resources assignment and timely vehicle scheduling are critical actions needed to organize successful secured operations that could, if well prepared, save many injured bodies and lessen the human distress. To reach this ultimate goal, Fateful choices must be taken; in some cases, they could be complex and critical such as hospital assignment, vehicle scheduling. Hence any failure in these procedures can potentially lead to increase mortality rate between wounded, thus a complete alertness and full caution should be exercised, Xin Ma et al. [2], Vahid Yadollahnejad et al. [3]. Generally, after the declaration of a crisis, a state of disorder appeared and resources are generally poorly distributed. The trajectories followed by ambulances from their depots to their destinations are indefinite and subject to several additional changes during the crisis, whether in the choice of paths to be covered, or in the choice of their destinations (wounded, hospitals ), this disorder leads to a waste of time and poor management of valuable resources (ambulances, hospitals) which increases our loss at several levels especially in terms of operating costs and humanitarian level (loss of life of the wounded) The general objective of our work is to solve the integrated problem of transport / allocation of wounded, in the most difficult cases, to minimize the human loss. The heart of our work is to present a strategy to address the following issues: 1. Serve wounded according to their priorities (scheduling according to their degrees of urgency). 2. Efficient assignment of ambulances to the wounded. 3. Efficient allocation of wounded to hospitals. 4. Minimal overall transport time of the entire operation (evacuation of all the injured). In fact, we treat firstly the Integrated Problem of Ambulance Scheduling and Resource Assignment (IPASRA) in the case of a sudden disaster in which we cannot identify neither disaster areas nor groups of wounded (FIG. 1). FIG. 1 – Multi-group Evacuation VS Individual Evacuation In a second part of our work, we studied the case of wounded transportation from multiple disaster areas to multiple hospitals (FIG. 1). In this part, our main objective is to solve the Integrated Problem of Ambulance Missions Scheduling and Resource Assignment (IPAMSRA). The main resources to be assigned are the ambulances and the hospitals. While, the hospitals serving capacities might be considered or not according to the extent of disaster and particularly to the number of wounded. We formulate both (IPASRA) and (IPAMSRA) as a linear model, furthermore a wellorganized hybrid algorithm grounded on GRASP and Tabu Search is designed to accomplish this objective. We will present numerical experiments illustrating the efficiency of our proposed algorithms and demonstrate that the generated solutions are of high-quality.