A Comprehensive Review on Scatter Search: Techniques, Applications, and Challenges

Table 5

Summary of scatter search’s variants.

Type	Algorithm	Description	Application	Performance	Ref.
Improved	ESSOL	Modification is done in reference set by using opposition-based learning scheme	Kinetic models of biochemical systems	Faster convergence rate, avoid local optima	[41]
	SS	Improvement in reference set	Optimization problems	Improvement in performance	[42]
	CSS	BLX-α operator, Solis and Wets algorithm, and Nelder–Mead simplex are employed to improve the performance of SS	Optimization problems	Better exploitation, better efficiency	[43]
Hybrid	OQNLP	Local NLP solver is employed to enhance the global search ability of SS	Optimization problems	Improvement in performance	[46]
	Bee-SS	Bee algorithm is integrated with SS to enhance the improvement method	Traveling salesman problem	More computational time than SS, better performance	[47]
	GA-SS	Crossover and mutation are used to replace combination and improvement	Knapsack problem	Better exploration capability	[48]
	CS-SS	CS is employed to enhance the improvement method	Traveling salesman problem	Better performance	[49]
	VNS-SS	VNS is used to replace both intensification and improvement components of SS to improve the quality solutions	Integer programming problem	Improvement in performance	[50]
	MCA-SS	MCA is used to update the reference set and improvement method of SS	Traveling salesman and job shop scheduling problems	Better exploration capability	[51]
	TS-SS	TS is used to modify the search element for better exploration and exploitation	Feedforward neural network problem	Better exploitation and exploration capability	[52]
	DR-SS	DR is employed to improve the search capability of SS	Mixed-integer programming	Improved performance	[53]
	eDE-SS	Enhanced DE is used to retain the quality solutions in long-term memory	Quadratic assignment and flow shop scheduling problems	Improve the effectiveness of SS	[54]
	DE-SS	Population reinitialization method is used to avoid the local optima	Optimization problems	Robust, better performance	[55]
	ACO-SS	ACO is incorporated in SS to improve the search ability	Multicast routing problems	Minimize the cost	[125]
Chaotic	ChSS	Four chaos maps are employed to generate CPRNG for SS	Flow shop problem	Better exploration capability	[56]
Binary	BSS	Modification in diversification generation, improvement, and combination methods	Knapsack problem, maximum diversity problem, max-cut problem	Improvement in performance	[57]
Multiobjective	AbYSS	Pareto dominance, density estimation, and external archive are incorporated in SS to store nondominated solutions	Optimization problems	Improve the performance	[58]
	SSMO	Ranking, crowding, and Pareto dominance are incorporated in SS	Optimization problems	Improve the performance	[59]
	MOSS	TS is used in the diversification generation method to generate Pareto-optimal solutions	Optimization problems	Better convergence, reasonable computational time	[60]
	SSPMO	TS is employed in SS to improve the Pareto-optimal solutions	Optimization problems	Better performance	[61]
	MPSOSS	A new leader selection is used to improve the convergence. PSO is incorporated in SS to improve spread	Optimization problems	Better convergence rate, improve spread	[62]
	RP-MOSS	A new preference-based TS and SS are introduced	Optimization problems	Better Pareto-optimal solutions	[63]
	MOSS-II	TS and convex combination methods are used as the diversification generation method. A constraint handling mechanism is employed to handle multiple constrains	Optimization problems	Improved performance	[64]
	MOSSSA	SA is employed as an improvement method. Fast archiving strategy is used to store the set of nondominated solutions	Water distribution network	Improved quality of network	[65]