1 package net.bmahe.genetics4j.core.spec.replacement;
2
3 /**
4 * Marker interface for replacement strategy specifications in evolutionary algorithms.
5 *
6 * <p>ReplacementStrategy defines the contract for specifying how offspring are integrated
7 * into the population to form the next generation. Replacement strategies determine which
8 * individuals survive from one generation to the next, balancing exploration of new solutions
9 * with preservation of good existing solutions.
10 *
11 * <p>Replacement strategies are used by:
12 * <ul>
13 * <li><strong>EA Configuration</strong>: To specify the survival strategy for generations</li>
14 * <li><strong>Replacement handlers</strong>: To create appropriate replacement implementations</li>
15 * <li><strong>Population management</strong>: To control population size and composition</li>
16 * <li><strong>Algorithm flow</strong>: To determine generational vs steady-state evolution</li>
17 * </ul>
18 *
19 * <p>The framework provides several concrete replacement strategy implementations:
20 * <ul>
21 * <li>{@link GenerationalReplacement}: Complete replacement of parent generation</li>
22 * <li>{@link Elitism}: Preserve best individuals while replacing others</li>
23 * <li>{@link DeleteNLast}: Remove worst N individuals from combined population</li>
24 * </ul>
25 *
26 * <p>Replacement strategies vary in their characteristics:
27 * <ul>
28 * <li><strong>Selection pressure</strong>: How aggressively poor solutions are eliminated</li>
29 * <li><strong>Diversity preservation</strong>: How well genetic diversity is maintained</li>
30 * <li><strong>Convergence speed</strong>: How quickly the algorithm converges to solutions</li>
31 * <li><strong>Population dynamics</strong>: How population composition changes over time</li>
32 * </ul>
33 *
34 * <p>Common replacement patterns:
35 * <ul>
36 * <li><strong>Generational</strong>: All parents replaced by offspring each generation</li>
37 * <li><strong>Steady-state</strong>: Only a few individuals replaced per iteration</li>
38 * <li><strong>Elitist</strong>: Best individuals always survive to next generation</li>
39 * <li><strong>Tournament replacement</strong>: Compete offspring against existing individuals</li>
40 * </ul>
41 *
42 * <p>Replacement strategy design considerations:
43 * <ul>
44 * <li><strong>Population size</strong>: Maintain constant population size across generations</li>
45 * <li><strong>Selection pressure</strong>: Balance exploitation with exploration</li>
46 * <li><strong>Premature convergence</strong>: Prevent loss of genetic diversity too early</li>
47 * <li><strong>Solution preservation</strong>: Ensure good solutions are not lost</li>
48 * </ul>
49 *
50 * <p>Interaction with other EA components:
51 * <ul>
52 * <li><strong>Selection policies</strong>: Work together to control evolutionary pressure</li>
53 * <li><strong>Offspring ratio</strong>: Number of offspring affects replacement decisions</li>
54 * <li><strong>Population size</strong>: Determines how many individuals can be replaced</li>
55 * <li><strong>Fitness evaluation</strong>: Required for fitness-based replacement decisions</li>
56 * </ul>
57 *
58 * <p>Example usage in genetic algorithm configuration:
59 * <pre>{@code
60 * // Generational replacement (all parents replaced)
61 * ReplacementStrategy generational = GenerationalReplacement.build();
62 *
63 * // Elitist replacement preserving top 10% of individuals
64 * ReplacementStrategy elitist = Elitism.builder()
65 * .offspringSelectionPolicy(Tournament.of(3))
66 * .offspringRatio(0.9) // Replace 90% of population
67 * .build();
68 *
69 * // Delete worst N individuals from combined population
70 * ReplacementStrategy deleteNLast = DeleteNLast.builder()
71 * .offspringSelectionPolicy(Tournament.of(2))
72 * .offspringRatio(0.5) // Generate 50% offspring
73 * .build();
74 *
75 * // Use in EA configuration
76 * EAConfiguration<Double> config = EAConfigurationBuilder.<Double>builder()
77 * .chromosomeSpecs(chromosomeSpec)
78 * .parentSelectionPolicy(Tournament.of(3))
79 * .replacementStrategy(elitist)
80 * .build();
81 * }</pre>
82 *
83 * <p>Performance implications:
84 * <ul>
85 * <li><strong>Computational cost</strong>: Some strategies require sorting or complex selection</li>
86 * <li><strong>Memory usage</strong>: Combined populations may temporarily increase memory needs</li>
87 * <li><strong>Convergence rate</strong>: Affects how quickly the algorithm finds good solutions</li>
88 * <li><strong>Solution quality</strong>: Influences the final quality of evolved solutions</li>
89 * </ul>
90 *
91 * @see net.bmahe.genetics4j.core.replacement.ReplacementStrategyImplementor
92 * @see net.bmahe.genetics4j.core.replacement.ReplacementStrategyHandler
93 * @see GenerationalReplacement
94 * @see Elitism
95 * @see DeleteNLast
96 */
97 public interface ReplacementStrategy {
98
99 }