Table 3.

Overview of the heuristics developed in this study for solving the MRRSP for emergency response.

Heuristic	Description	Related Works	Shortcomings
Greedy Algorithm	–Decision rule is used to prioritise demand points (nodes) to visit	Liu et al. [65]; Ceselli et al. [66]; Majzoubi [67]; Tang and Zhu [68]; Zhao et al. [69]; Ciancio et al. [70]	–Solution may not be the best especially when there are multiple factors to consider
	–Optimal nodes are then selected to constitute the route or schedule of the vehicle		–Can be trapped in a local optimum instead of finding the global optimum
Augmented Greedy	–Adjustments are made to the priorities of the nodes based on the outcome of the Greedy algorithm	Li and Wang [71]; Almutairi [72]; Bettinelli et al. [73]; Kritikos and Ioannou [74]	–Same issues with the Greedy algorithm except that the final solutions might be improved
Algorithm	–Rerouting or rescheduling is then initiated
k-Node Crossover Algorithm	–Crossover procedure is applied to improve initial solution of a heuristic	Baptista and Tavares [75]; Zhang et al. [76]; Zheng et al. [77]	–Longer running time due to many iterations during the crossover
	–A certain number of nodes in the route are randomly exchanged prior to recalculation
Scheduling Algorithm	–More sophisticated decision rule taking multiple factors into account is used to prioritise nodes to visit	Jaw et al. [78]; Weng et al. [79]; Ramchurn et al. [80]; Wex et al. [15]	–Can be trapped in a local optimum instead of finding the global optimum
	–Adjustments are made to the priorities of the nodes based on the outcome
Monte Carlo Algorithm	–Once the nodes are prioritised, a certain percentage of them are randomised to generate a route or schedule	Wex et al. [81]; Abdullah et al. [82]; Al-Harthei et al. [83]; Wu and Sioshansi [84]	–Can be challenging to determine the appropriate level of randomness and number of iterations
	–Instead of a single iteration, multiple iterations are used to find near-optimal solutions		–Running time can be long due to many iterations
Genetic Algorithm	–A meta-heuristic served as a benchmark for assessing the performance of other heuristics	Baker and Ayechew [85]; Okhrin and Richter [86]; Zidi et al. [87]; Mguis et al. [88]; Zheng et al. [77]; Qin et al. [89]	–Running time can be long due to many iterations
	–Based on the principle of evolution with crossover of chromosomes, representing a sequence of nodes in a route, to find better solutions
Clustering Algorithm	–A large problem is first broken down into a number of sub-problems, each with many clusters, and solved using the exact method or heuristics	Özdamar and Demir [90]; He et al. [91]; Vargas-Florez et al. [92]; Pillac et al. [17]; Gharib et al. [93]; Penna et al. [94]	–Can be challenging in splitting the original problem into an appropriate number of clusters to obtain optimality
	–Solutions for the sub-problems are then aggregated to form the overall solution of the bigger problem		–Running time can be long due to many iterations